Training in progress, epoch 0
Browse files- adapter_config.json +21 -0
- adapter_model.bin +3 -0
- checkpoint-3500/README.md +20 -0
- checkpoint-3500/adapter_config.json +21 -0
- checkpoint-3500/adapter_model.bin +3 -0
- checkpoint-3500/optimizer.pt +3 -0
- checkpoint-3500/rng_state.pth +3 -0
- checkpoint-3500/scheduler.pt +3 -0
- checkpoint-3500/special_tokens_map.json +24 -0
- checkpoint-3500/tokenizer.json +0 -0
- checkpoint-3500/tokenizer_config.json +32 -0
- checkpoint-3500/trainer_state.json +2116 -0
- checkpoint-3500/training_args.bin +3 -0
- runs/Nov24_18-45-12_0cb8a69e401c/events.out.tfevents.1700851546.0cb8a69e401c.2133.0 +3 -0
- runs/Nov24_18-46-44_0cb8a69e401c/events.out.tfevents.1700851610.0cb8a69e401c.2133.1 +3 -0
- runs/Nov24_18-59-08_0cb8a69e401c/events.out.tfevents.1700852369.0cb8a69e401c.31161.0 +3 -0
- runs/Nov25_00-57-56_0cb8a69e401c/events.out.tfevents.1700873879.0cb8a69e401c.31161.1 +3 -0
- runs/Nov25_01-00-09_0cb8a69e401c/events.out.tfevents.1700874012.0cb8a69e401c.31161.2 +3 -0
- special_tokens_map.json +24 -0
- tokenizer.json +0 -0
- tokenizer_config.json +32 -0
- training_args.bin +3 -0
adapter_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "NousResearch/Nous-Hermes-Llama2-13b",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 32,
|
| 11 |
+
"lora_dropout": 0.1,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"v_proj"
|
| 19 |
+
],
|
| 20 |
+
"task_type": "CAUSAL_LM"
|
| 21 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:939438a277b9f98af60018a9d898efc87f07cce8186a4140fe7b4efc11d9480d
|
| 3 |
+
size 52486922
|
checkpoint-3500/README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- load_in_8bit: False
|
| 9 |
+
- load_in_4bit: True
|
| 10 |
+
- llm_int8_threshold: 6.0
|
| 11 |
+
- llm_int8_skip_modules: None
|
| 12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 13 |
+
- llm_int8_has_fp16_weight: False
|
| 14 |
+
- bnb_4bit_quant_type: nf4
|
| 15 |
+
- bnb_4bit_use_double_quant: True
|
| 16 |
+
- bnb_4bit_compute_dtype: bfloat16
|
| 17 |
+
### Framework versions
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
- PEFT 0.4.0
|
checkpoint-3500/adapter_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "NousResearch/Nous-Hermes-Llama2-13b",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 32,
|
| 11 |
+
"lora_dropout": 0.1,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 16,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"v_proj"
|
| 19 |
+
],
|
| 20 |
+
"task_type": "CAUSAL_LM"
|
| 21 |
+
}
|
checkpoint-3500/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:939438a277b9f98af60018a9d898efc87f07cce8186a4140fe7b4efc11d9480d
|
| 3 |
+
size 52486922
|
checkpoint-3500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:39a60f8ed28e96b99770244bb1506b1e94eedc4b9fb442898783963af42991ea
|
| 3 |
+
size 104992698
|
checkpoint-3500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:900e99e3799bf608b0f62dd23536243ff7dbe6316cb1b4a988e666a78d2098ee
|
| 3 |
+
size 14244
|
checkpoint-3500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:de3d3c5cfa9fff6809c38b1bced0aee3ba50cbe6d3fcb178157361c7bc12b382
|
| 3 |
+
size 1064
|
checkpoint-3500/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<unk>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": true,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-3500/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-3500/tokenizer_config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"__type": "AddedToken",
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"clean_up_tokenization_spaces": false,
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "</s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"legacy": false,
|
| 20 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"sp_model_kwargs": {},
|
| 23 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 24 |
+
"unk_token": {
|
| 25 |
+
"__type": "AddedToken",
|
| 26 |
+
"content": "<unk>",
|
| 27 |
+
"lstrip": false,
|
| 28 |
+
"normalized": true,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"single_word": false
|
| 31 |
+
}
|
| 32 |
+
}
|
checkpoint-3500/trainer_state.json
ADDED
|
@@ -0,0 +1,2116 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.24560541735377706,
|
| 5 |
+
"global_step": 3500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 1.1428571428571429e-05,
|
| 13 |
+
"loss": 1.2994,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 2.2857142857142858e-05,
|
| 19 |
+
"loss": 1.2625,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.0,
|
| 24 |
+
"learning_rate": 3.428571428571429e-05,
|
| 25 |
+
"loss": 1.1582,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.0,
|
| 30 |
+
"learning_rate": 4.5714285714285716e-05,
|
| 31 |
+
"loss": 0.9934,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.0,
|
| 36 |
+
"learning_rate": 5.714285714285714e-05,
|
| 37 |
+
"loss": 0.8862,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.0,
|
| 42 |
+
"learning_rate": 6.857142857142858e-05,
|
| 43 |
+
"loss": 0.7613,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0,
|
| 48 |
+
"learning_rate": 8e-05,
|
| 49 |
+
"loss": 0.6832,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.01,
|
| 54 |
+
"learning_rate": 9.142857142857143e-05,
|
| 55 |
+
"loss": 0.5934,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.01,
|
| 60 |
+
"learning_rate": 0.00010285714285714286,
|
| 61 |
+
"loss": 0.5305,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.01,
|
| 66 |
+
"learning_rate": 0.00011428571428571428,
|
| 67 |
+
"loss": 0.5195,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.01,
|
| 72 |
+
"learning_rate": 0.00012571428571428572,
|
| 73 |
+
"loss": 0.5001,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.01,
|
| 78 |
+
"learning_rate": 0.00013714285714285716,
|
| 79 |
+
"loss": 0.4812,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.01,
|
| 84 |
+
"learning_rate": 0.00014857142857142857,
|
| 85 |
+
"loss": 0.4624,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.01,
|
| 90 |
+
"learning_rate": 0.00016,
|
| 91 |
+
"loss": 0.4699,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.01,
|
| 96 |
+
"learning_rate": 0.00017142857142857143,
|
| 97 |
+
"loss": 0.4355,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.01,
|
| 102 |
+
"learning_rate": 0.00018285714285714286,
|
| 103 |
+
"loss": 0.4341,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.01,
|
| 108 |
+
"learning_rate": 0.00019314285714285717,
|
| 109 |
+
"loss": 0.4243,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.01,
|
| 114 |
+
"learning_rate": 0.00020457142857142858,
|
| 115 |
+
"loss": 0.4255,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.01,
|
| 120 |
+
"learning_rate": 0.00021600000000000002,
|
| 121 |
+
"loss": 0.4116,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.01,
|
| 126 |
+
"learning_rate": 0.00022742857142857146,
|
| 127 |
+
"loss": 0.425,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.01,
|
| 132 |
+
"learning_rate": 0.0002388571428571429,
|
| 133 |
+
"loss": 0.4121,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.02,
|
| 138 |
+
"learning_rate": 0.0002502857142857143,
|
| 139 |
+
"loss": 0.4066,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.02,
|
| 144 |
+
"learning_rate": 0.0002617142857142857,
|
| 145 |
+
"loss": 0.4019,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.02,
|
| 150 |
+
"learning_rate": 0.00027314285714285716,
|
| 151 |
+
"loss": 0.3961,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.02,
|
| 156 |
+
"learning_rate": 0.00028457142857142857,
|
| 157 |
+
"loss": 0.3947,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.02,
|
| 162 |
+
"learning_rate": 0.000296,
|
| 163 |
+
"loss": 0.4015,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.02,
|
| 168 |
+
"learning_rate": 0.00030742857142857145,
|
| 169 |
+
"loss": 0.3862,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.02,
|
| 174 |
+
"learning_rate": 0.00031885714285714286,
|
| 175 |
+
"loss": 0.4013,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.02,
|
| 180 |
+
"learning_rate": 0.00033028571428571433,
|
| 181 |
+
"loss": 0.3913,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.02,
|
| 186 |
+
"learning_rate": 0.00034171428571428574,
|
| 187 |
+
"loss": 0.4038,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.02,
|
| 192 |
+
"learning_rate": 0.00035314285714285715,
|
| 193 |
+
"loss": 0.3962,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.02,
|
| 198 |
+
"learning_rate": 0.0003645714285714286,
|
| 199 |
+
"loss": 0.3897,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.02,
|
| 204 |
+
"learning_rate": 0.000376,
|
| 205 |
+
"loss": 0.3875,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.02,
|
| 210 |
+
"learning_rate": 0.00038742857142857144,
|
| 211 |
+
"loss": 0.378,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.02,
|
| 216 |
+
"learning_rate": 0.00039885714285714286,
|
| 217 |
+
"loss": 0.3843,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.03,
|
| 222 |
+
"learning_rate": 0.00039999194323417424,
|
| 223 |
+
"loss": 0.3827,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.03,
|
| 228 |
+
"learning_rate": 0.000399964093519005,
|
| 229 |
+
"loss": 0.3828,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.03,
|
| 234 |
+
"learning_rate": 0.0003999163541576902,
|
| 235 |
+
"loss": 0.3905,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.03,
|
| 240 |
+
"learning_rate": 0.00039984872989867726,
|
| 241 |
+
"loss": 0.3938,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.03,
|
| 246 |
+
"learning_rate": 0.0003997612274682868,
|
| 247 |
+
"loss": 0.3811,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.03,
|
| 252 |
+
"learning_rate": 0.0003996538555700433,
|
| 253 |
+
"loss": 0.3942,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.03,
|
| 258 |
+
"learning_rate": 0.0003995266248838093,
|
| 259 |
+
"loss": 0.3746,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.03,
|
| 264 |
+
"learning_rate": 0.00039937954806472354,
|
| 265 |
+
"loss": 0.3785,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.03,
|
| 270 |
+
"learning_rate": 0.0003992126397419419,
|
| 271 |
+
"loss": 0.3807,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.03,
|
| 276 |
+
"learning_rate": 0.00039902591651718225,
|
| 277 |
+
"loss": 0.3765,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.03,
|
| 282 |
+
"learning_rate": 0.0003988193969630734,
|
| 283 |
+
"loss": 0.375,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.03,
|
| 288 |
+
"learning_rate": 0.0003985931016213074,
|
| 289 |
+
"loss": 0.3886,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.03,
|
| 294 |
+
"learning_rate": 0.00039834705300059676,
|
| 295 |
+
"loss": 0.3796,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.03,
|
| 300 |
+
"learning_rate": 0.0003980812755744352,
|
| 301 |
+
"loss": 0.3778,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.04,
|
| 306 |
+
"learning_rate": 0.00039779579577866366,
|
| 307 |
+
"loss": 0.3801,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.04,
|
| 312 |
+
"learning_rate": 0.0003974906420088406,
|
| 313 |
+
"loss": 0.382,
|
| 314 |
+
"step": 510
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.04,
|
| 318 |
+
"learning_rate": 0.00039716584461741783,
|
| 319 |
+
"loss": 0.3779,
|
| 320 |
+
"step": 520
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.04,
|
| 324 |
+
"learning_rate": 0.0003968214359107213,
|
| 325 |
+
"loss": 0.3811,
|
| 326 |
+
"step": 530
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.04,
|
| 330 |
+
"learning_rate": 0.0003964574501457378,
|
| 331 |
+
"loss": 0.3837,
|
| 332 |
+
"step": 540
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.04,
|
| 336 |
+
"learning_rate": 0.0003960739235267074,
|
| 337 |
+
"loss": 0.3798,
|
| 338 |
+
"step": 550
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.04,
|
| 342 |
+
"learning_rate": 0.00039567089420152265,
|
| 343 |
+
"loss": 0.3801,
|
| 344 |
+
"step": 560
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.04,
|
| 348 |
+
"learning_rate": 0.00039524840225793395,
|
| 349 |
+
"loss": 0.3761,
|
| 350 |
+
"step": 570
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.04,
|
| 354 |
+
"learning_rate": 0.00039480648971956214,
|
| 355 |
+
"loss": 0.3702,
|
| 356 |
+
"step": 580
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.04,
|
| 360 |
+
"learning_rate": 0.00039434520054171875,
|
| 361 |
+
"loss": 0.3787,
|
| 362 |
+
"step": 590
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.04,
|
| 366 |
+
"learning_rate": 0.0003938645806070337,
|
| 367 |
+
"loss": 0.364,
|
| 368 |
+
"step": 600
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.04,
|
| 372 |
+
"learning_rate": 0.00039336467772089194,
|
| 373 |
+
"loss": 0.3669,
|
| 374 |
+
"step": 610
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.04,
|
| 378 |
+
"learning_rate": 0.0003928455416066777,
|
| 379 |
+
"loss": 0.3713,
|
| 380 |
+
"step": 620
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.04,
|
| 384 |
+
"learning_rate": 0.00039230722390082984,
|
| 385 |
+
"loss": 0.3676,
|
| 386 |
+
"step": 630
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.04,
|
| 390 |
+
"learning_rate": 0.0003917497781477043,
|
| 391 |
+
"loss": 0.3715,
|
| 392 |
+
"step": 640
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.05,
|
| 396 |
+
"learning_rate": 0.00039117325979424984,
|
| 397 |
+
"loss": 0.3626,
|
| 398 |
+
"step": 650
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.05,
|
| 402 |
+
"learning_rate": 0.00039057772618449163,
|
| 403 |
+
"loss": 0.3744,
|
| 404 |
+
"step": 660
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.05,
|
| 408 |
+
"learning_rate": 0.0003899632365538282,
|
| 409 |
+
"loss": 0.3686,
|
| 410 |
+
"step": 670
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.05,
|
| 414 |
+
"learning_rate": 0.0003893298520231393,
|
| 415 |
+
"loss": 0.3683,
|
| 416 |
+
"step": 680
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.05,
|
| 420 |
+
"learning_rate": 0.0003886776355927065,
|
| 421 |
+
"loss": 0.3745,
|
| 422 |
+
"step": 690
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.05,
|
| 426 |
+
"learning_rate": 0.00038800665213594663,
|
| 427 |
+
"loss": 0.3701,
|
| 428 |
+
"step": 700
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.05,
|
| 432 |
+
"learning_rate": 0.0003873169683929594,
|
| 433 |
+
"loss": 0.3685,
|
| 434 |
+
"step": 710
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.05,
|
| 438 |
+
"learning_rate": 0.0003866086529638889,
|
| 439 |
+
"loss": 0.3649,
|
| 440 |
+
"step": 720
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.05,
|
| 444 |
+
"learning_rate": 0.0003858817763020997,
|
| 445 |
+
"loss": 0.3621,
|
| 446 |
+
"step": 730
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.05,
|
| 450 |
+
"learning_rate": 0.00038513641070717016,
|
| 451 |
+
"loss": 0.3678,
|
| 452 |
+
"step": 740
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.05,
|
| 456 |
+
"learning_rate": 0.0003843726303177001,
|
| 457 |
+
"loss": 0.362,
|
| 458 |
+
"step": 750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.05,
|
| 462 |
+
"learning_rate": 0.0003835905111039371,
|
| 463 |
+
"loss": 0.3551,
|
| 464 |
+
"step": 760
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.05,
|
| 468 |
+
"learning_rate": 0.00038279013086021973,
|
| 469 |
+
"loss": 0.3675,
|
| 470 |
+
"step": 770
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.05,
|
| 474 |
+
"learning_rate": 0.00038197156919723984,
|
| 475 |
+
"loss": 0.3666,
|
| 476 |
+
"step": 780
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.06,
|
| 480 |
+
"learning_rate": 0.000381134907534124,
|
| 481 |
+
"loss": 0.3719,
|
| 482 |
+
"step": 790
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.06,
|
| 486 |
+
"learning_rate": 0.0003802802290903348,
|
| 487 |
+
"loss": 0.3675,
|
| 488 |
+
"step": 800
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.06,
|
| 492 |
+
"learning_rate": 0.00037940761887739373,
|
| 493 |
+
"loss": 0.3746,
|
| 494 |
+
"step": 810
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.06,
|
| 498 |
+
"learning_rate": 0.00037851716369042513,
|
| 499 |
+
"loss": 0.3677,
|
| 500 |
+
"step": 820
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.06,
|
| 504 |
+
"learning_rate": 0.000377608952099523,
|
| 505 |
+
"loss": 0.3689,
|
| 506 |
+
"step": 830
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.06,
|
| 510 |
+
"learning_rate": 0.00037668307444094153,
|
| 511 |
+
"loss": 0.3606,
|
| 512 |
+
"step": 840
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.06,
|
| 516 |
+
"learning_rate": 0.00037573962280810944,
|
| 517 |
+
"loss": 0.3657,
|
| 518 |
+
"step": 850
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.06,
|
| 522 |
+
"learning_rate": 0.0003747786910424701,
|
| 523 |
+
"loss": 0.3601,
|
| 524 |
+
"step": 860
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.06,
|
| 528 |
+
"learning_rate": 0.00037380037472414704,
|
| 529 |
+
"loss": 0.3732,
|
| 530 |
+
"step": 870
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.06,
|
| 534 |
+
"learning_rate": 0.0003728047711624374,
|
| 535 |
+
"loss": 0.3765,
|
| 536 |
+
"step": 880
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.06,
|
| 540 |
+
"learning_rate": 0.0003717919793861329,
|
| 541 |
+
"loss": 0.3641,
|
| 542 |
+
"step": 890
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.06,
|
| 546 |
+
"learning_rate": 0.00037076210013366943,
|
| 547 |
+
"loss": 0.3621,
|
| 548 |
+
"step": 900
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.06,
|
| 552 |
+
"learning_rate": 0.00036971523584310764,
|
| 553 |
+
"loss": 0.3635,
|
| 554 |
+
"step": 910
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.06,
|
| 558 |
+
"learning_rate": 0.0003686514906419431,
|
| 559 |
+
"loss": 0.3716,
|
| 560 |
+
"step": 920
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.07,
|
| 564 |
+
"learning_rate": 0.0003675709703367499,
|
| 565 |
+
"loss": 0.3613,
|
| 566 |
+
"step": 930
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.07,
|
| 570 |
+
"learning_rate": 0.0003664737824026558,
|
| 571 |
+
"loss": 0.3728,
|
| 572 |
+
"step": 940
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.07,
|
| 576 |
+
"learning_rate": 0.0003653600359726526,
|
| 577 |
+
"loss": 0.362,
|
| 578 |
+
"step": 950
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.07,
|
| 582 |
+
"learning_rate": 0.00036422984182674084,
|
| 583 |
+
"loss": 0.3518,
|
| 584 |
+
"step": 960
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.07,
|
| 588 |
+
"learning_rate": 0.00036308331238091095,
|
| 589 |
+
"loss": 0.3672,
|
| 590 |
+
"step": 970
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.07,
|
| 594 |
+
"learning_rate": 0.00036192056167596193,
|
| 595 |
+
"loss": 0.3563,
|
| 596 |
+
"step": 980
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.07,
|
| 600 |
+
"learning_rate": 0.00036074170536615777,
|
| 601 |
+
"loss": 0.358,
|
| 602 |
+
"step": 990
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.07,
|
| 606 |
+
"learning_rate": 0.00035954686070772404,
|
| 607 |
+
"loss": 0.3594,
|
| 608 |
+
"step": 1000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.07,
|
| 612 |
+
"learning_rate": 0.00035833614654718493,
|
| 613 |
+
"loss": 0.3635,
|
| 614 |
+
"step": 1010
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.07,
|
| 618 |
+
"learning_rate": 0.00035710968330954164,
|
| 619 |
+
"loss": 0.354,
|
| 620 |
+
"step": 1020
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.07,
|
| 624 |
+
"learning_rate": 0.0003558675929862949,
|
| 625 |
+
"loss": 0.3721,
|
| 626 |
+
"step": 1030
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.07,
|
| 630 |
+
"learning_rate": 0.00035460999912330994,
|
| 631 |
+
"loss": 0.3588,
|
| 632 |
+
"step": 1040
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.07,
|
| 636 |
+
"learning_rate": 0.0003533370268085289,
|
| 637 |
+
"loss": 0.3608,
|
| 638 |
+
"step": 1050
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.07,
|
| 642 |
+
"learning_rate": 0.0003520488026595282,
|
| 643 |
+
"loss": 0.3539,
|
| 644 |
+
"step": 1060
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.08,
|
| 648 |
+
"learning_rate": 0.00035074545481092456,
|
| 649 |
+
"loss": 0.3597,
|
| 650 |
+
"step": 1070
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.08,
|
| 654 |
+
"learning_rate": 0.00034942711290162997,
|
| 655 |
+
"loss": 0.3493,
|
| 656 |
+
"step": 1080
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.08,
|
| 660 |
+
"learning_rate": 0.00034809390806195685,
|
| 661 |
+
"loss": 0.3648,
|
| 662 |
+
"step": 1090
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.08,
|
| 666 |
+
"learning_rate": 0.0003467459729005753,
|
| 667 |
+
"loss": 0.3553,
|
| 668 |
+
"step": 1100
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.08,
|
| 672 |
+
"learning_rate": 0.0003453834414913229,
|
| 673 |
+
"loss": 0.3497,
|
| 674 |
+
"step": 1110
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.08,
|
| 678 |
+
"learning_rate": 0.00034400644935986876,
|
| 679 |
+
"loss": 0.3496,
|
| 680 |
+
"step": 1120
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.08,
|
| 684 |
+
"learning_rate": 0.0003426151334702336,
|
| 685 |
+
"loss": 0.3612,
|
| 686 |
+
"step": 1130
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.08,
|
| 690 |
+
"learning_rate": 0.00034120963221116637,
|
| 691 |
+
"loss": 0.3578,
|
| 692 |
+
"step": 1140
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.08,
|
| 696 |
+
"learning_rate": 0.00033979008538237914,
|
| 697 |
+
"loss": 0.3502,
|
| 698 |
+
"step": 1150
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.08,
|
| 702 |
+
"learning_rate": 0.0003383566341806417,
|
| 703 |
+
"loss": 0.3556,
|
| 704 |
+
"step": 1160
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.08,
|
| 708 |
+
"learning_rate": 0.00033690942118573774,
|
| 709 |
+
"loss": 0.3525,
|
| 710 |
+
"step": 1170
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.08,
|
| 714 |
+
"learning_rate": 0.0003354485903462825,
|
| 715 |
+
"loss": 0.3524,
|
| 716 |
+
"step": 1180
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.08,
|
| 720 |
+
"learning_rate": 0.000333974286965405,
|
| 721 |
+
"loss": 0.353,
|
| 722 |
+
"step": 1190
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.08,
|
| 726 |
+
"learning_rate": 0.0003324866576862952,
|
| 727 |
+
"loss": 0.362,
|
| 728 |
+
"step": 1200
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.08,
|
| 732 |
+
"learning_rate": 0.00033098585047761797,
|
| 733 |
+
"loss": 0.3613,
|
| 734 |
+
"step": 1210
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.09,
|
| 738 |
+
"learning_rate": 0.0003294720146187955,
|
| 739 |
+
"loss": 0.3514,
|
| 740 |
+
"step": 1220
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.09,
|
| 744 |
+
"learning_rate": 0.00032794530068515874,
|
| 745 |
+
"loss": 0.3638,
|
| 746 |
+
"step": 1230
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.09,
|
| 750 |
+
"learning_rate": 0.0003264058605329702,
|
| 751 |
+
"loss": 0.3511,
|
| 752 |
+
"step": 1240
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.09,
|
| 756 |
+
"learning_rate": 0.0003248538472843198,
|
| 757 |
+
"loss": 0.3619,
|
| 758 |
+
"step": 1250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.09,
|
| 762 |
+
"learning_rate": 0.00032328941531189397,
|
| 763 |
+
"loss": 0.3535,
|
| 764 |
+
"step": 1260
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.09,
|
| 768 |
+
"learning_rate": 0.00032171272022362134,
|
| 769 |
+
"loss": 0.3523,
|
| 770 |
+
"step": 1270
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.09,
|
| 774 |
+
"learning_rate": 0.0003201239188471943,
|
| 775 |
+
"loss": 0.3581,
|
| 776 |
+
"step": 1280
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.09,
|
| 780 |
+
"learning_rate": 0.0003185231692144706,
|
| 781 |
+
"loss": 0.3555,
|
| 782 |
+
"step": 1290
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.09,
|
| 786 |
+
"learning_rate": 0.00031691063054575434,
|
| 787 |
+
"loss": 0.3472,
|
| 788 |
+
"step": 1300
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.09,
|
| 792 |
+
"learning_rate": 0.0003152864632339587,
|
| 793 |
+
"loss": 0.3538,
|
| 794 |
+
"step": 1310
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.09,
|
| 798 |
+
"learning_rate": 0.0003136508288286527,
|
| 799 |
+
"loss": 0.3477,
|
| 800 |
+
"step": 1320
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.09,
|
| 804 |
+
"learning_rate": 0.0003120038900199922,
|
| 805 |
+
"loss": 0.3475,
|
| 806 |
+
"step": 1330
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.09,
|
| 810 |
+
"learning_rate": 0.00031034581062253786,
|
| 811 |
+
"loss": 0.3521,
|
| 812 |
+
"step": 1340
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.09,
|
| 816 |
+
"learning_rate": 0.0003086767555589612,
|
| 817 |
+
"loss": 0.3531,
|
| 818 |
+
"step": 1350
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.1,
|
| 822 |
+
"learning_rate": 0.0003069968908436401,
|
| 823 |
+
"loss": 0.3448,
|
| 824 |
+
"step": 1360
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.1,
|
| 828 |
+
"learning_rate": 0.00030530638356614644,
|
| 829 |
+
"loss": 0.3564,
|
| 830 |
+
"step": 1370
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.1,
|
| 834 |
+
"learning_rate": 0.00030360540187462606,
|
| 835 |
+
"loss": 0.3473,
|
| 836 |
+
"step": 1380
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.1,
|
| 840 |
+
"learning_rate": 0.00030189411495907376,
|
| 841 |
+
"loss": 0.3575,
|
| 842 |
+
"step": 1390
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.1,
|
| 846 |
+
"learning_rate": 0.0003001726930345046,
|
| 847 |
+
"loss": 0.3478,
|
| 848 |
+
"step": 1400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.1,
|
| 852 |
+
"learning_rate": 0.00029844130732402347,
|
| 853 |
+
"loss": 0.3499,
|
| 854 |
+
"step": 1410
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.1,
|
| 858 |
+
"learning_rate": 0.000296700130041794,
|
| 859 |
+
"loss": 0.3638,
|
| 860 |
+
"step": 1420
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.1,
|
| 864 |
+
"learning_rate": 0.00029494933437590926,
|
| 865 |
+
"loss": 0.3494,
|
| 866 |
+
"step": 1430
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.1,
|
| 870 |
+
"learning_rate": 0.00029318909447116496,
|
| 871 |
+
"loss": 0.3497,
|
| 872 |
+
"step": 1440
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.1,
|
| 876 |
+
"learning_rate": 0.0002914195854117389,
|
| 877 |
+
"loss": 0.3555,
|
| 878 |
+
"step": 1450
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.1,
|
| 882 |
+
"learning_rate": 0.0002896409832037748,
|
| 883 |
+
"loss": 0.352,
|
| 884 |
+
"step": 1460
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.1,
|
| 888 |
+
"learning_rate": 0.0002878534647578768,
|
| 889 |
+
"loss": 0.3486,
|
| 890 |
+
"step": 1470
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.1,
|
| 894 |
+
"learning_rate": 0.0002860572078715121,
|
| 895 |
+
"loss": 0.3528,
|
| 896 |
+
"step": 1480
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.1,
|
| 900 |
+
"learning_rate": 0.0002842523912113264,
|
| 901 |
+
"loss": 0.3412,
|
| 902 |
+
"step": 1490
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.11,
|
| 906 |
+
"learning_rate": 0.00028243919429537265,
|
| 907 |
+
"loss": 0.3384,
|
| 908 |
+
"step": 1500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.11,
|
| 912 |
+
"learning_rate": 0.00028061779747525504,
|
| 913 |
+
"loss": 0.3493,
|
| 914 |
+
"step": 1510
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.11,
|
| 918 |
+
"learning_rate": 0.00027878838191819024,
|
| 919 |
+
"loss": 0.3416,
|
| 920 |
+
"step": 1520
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.11,
|
| 924 |
+
"learning_rate": 0.0002769511295889872,
|
| 925 |
+
"loss": 0.3485,
|
| 926 |
+
"step": 1530
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.11,
|
| 930 |
+
"learning_rate": 0.000275106223231948,
|
| 931 |
+
"loss": 0.3436,
|
| 932 |
+
"step": 1540
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.11,
|
| 936 |
+
"learning_rate": 0.00027325384635269094,
|
| 937 |
+
"loss": 0.3405,
|
| 938 |
+
"step": 1550
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.11,
|
| 942 |
+
"learning_rate": 0.00027139418319989785,
|
| 943 |
+
"loss": 0.349,
|
| 944 |
+
"step": 1560
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.11,
|
| 948 |
+
"learning_rate": 0.0002695274187469878,
|
| 949 |
+
"loss": 0.3366,
|
| 950 |
+
"step": 1570
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.11,
|
| 954 |
+
"learning_rate": 0.0002676537386737183,
|
| 955 |
+
"loss": 0.3524,
|
| 956 |
+
"step": 1580
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.11,
|
| 960 |
+
"learning_rate": 0.00026577332934771665,
|
| 961 |
+
"loss": 0.3405,
|
| 962 |
+
"step": 1590
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.11,
|
| 966 |
+
"learning_rate": 0.0002638863778059426,
|
| 967 |
+
"loss": 0.3437,
|
| 968 |
+
"step": 1600
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.11,
|
| 972 |
+
"learning_rate": 0.00026199307173608433,
|
| 973 |
+
"loss": 0.3375,
|
| 974 |
+
"step": 1610
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.11,
|
| 978 |
+
"learning_rate": 0.0002600935994578904,
|
| 979 |
+
"loss": 0.3447,
|
| 980 |
+
"step": 1620
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.11,
|
| 984 |
+
"learning_rate": 0.0002581881499044377,
|
| 985 |
+
"loss": 0.3464,
|
| 986 |
+
"step": 1630
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.12,
|
| 990 |
+
"learning_rate": 0.0002562769126033394,
|
| 991 |
+
"loss": 0.3369,
|
| 992 |
+
"step": 1640
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.12,
|
| 996 |
+
"learning_rate": 0.00025436007765789327,
|
| 997 |
+
"loss": 0.3423,
|
| 998 |
+
"step": 1650
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.12,
|
| 1002 |
+
"learning_rate": 0.00025243783572817297,
|
| 1003 |
+
"loss": 0.3433,
|
| 1004 |
+
"step": 1660
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.12,
|
| 1008 |
+
"learning_rate": 0.0002505103780120636,
|
| 1009 |
+
"loss": 0.3412,
|
| 1010 |
+
"step": 1670
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.12,
|
| 1014 |
+
"learning_rate": 0.0002485778962262443,
|
| 1015 |
+
"loss": 0.3424,
|
| 1016 |
+
"step": 1680
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.12,
|
| 1020 |
+
"learning_rate": 0.00024664058258711853,
|
| 1021 |
+
"loss": 0.3452,
|
| 1022 |
+
"step": 1690
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.12,
|
| 1026 |
+
"learning_rate": 0.0002446986297916954,
|
| 1027 |
+
"loss": 0.3493,
|
| 1028 |
+
"step": 1700
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.12,
|
| 1032 |
+
"learning_rate": 0.00024275223099842291,
|
| 1033 |
+
"loss": 0.3387,
|
| 1034 |
+
"step": 1710
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.12,
|
| 1038 |
+
"learning_rate": 0.00024080157980797484,
|
| 1039 |
+
"loss": 0.3421,
|
| 1040 |
+
"step": 1720
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.12,
|
| 1044 |
+
"learning_rate": 0.0002388468702439944,
|
| 1045 |
+
"loss": 0.3365,
|
| 1046 |
+
"step": 1730
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.12,
|
| 1050 |
+
"learning_rate": 0.00023688829673379534,
|
| 1051 |
+
"loss": 0.3382,
|
| 1052 |
+
"step": 1740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.12,
|
| 1056 |
+
"learning_rate": 0.00023492605408902297,
|
| 1057 |
+
"loss": 0.3443,
|
| 1058 |
+
"step": 1750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.12,
|
| 1062 |
+
"learning_rate": 0.00023296033748627712,
|
| 1063 |
+
"loss": 0.3441,
|
| 1064 |
+
"step": 1760
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.12,
|
| 1068 |
+
"learning_rate": 0.0002309913424476986,
|
| 1069 |
+
"loss": 0.3425,
|
| 1070 |
+
"step": 1770
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.12,
|
| 1074 |
+
"learning_rate": 0.00022901926482152138,
|
| 1075 |
+
"loss": 0.3469,
|
| 1076 |
+
"step": 1780
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.13,
|
| 1080 |
+
"learning_rate": 0.00022704430076259246,
|
| 1081 |
+
"loss": 0.3409,
|
| 1082 |
+
"step": 1790
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.13,
|
| 1086 |
+
"learning_rate": 0.00022506664671286087,
|
| 1087 |
+
"loss": 0.3585,
|
| 1088 |
+
"step": 1800
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.13,
|
| 1092 |
+
"learning_rate": 0.00022308649938183864,
|
| 1093 |
+
"loss": 0.3327,
|
| 1094 |
+
"step": 1810
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.13,
|
| 1098 |
+
"learning_rate": 0.00022110405572703466,
|
| 1099 |
+
"loss": 0.3354,
|
| 1100 |
+
"step": 1820
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.13,
|
| 1104 |
+
"learning_rate": 0.00021911951293436416,
|
| 1105 |
+
"loss": 0.3424,
|
| 1106 |
+
"step": 1830
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.13,
|
| 1110 |
+
"learning_rate": 0.00021713306839853545,
|
| 1111 |
+
"loss": 0.3369,
|
| 1112 |
+
"step": 1840
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.13,
|
| 1116 |
+
"learning_rate": 0.0002151449197034157,
|
| 1117 |
+
"loss": 0.3407,
|
| 1118 |
+
"step": 1850
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.13,
|
| 1122 |
+
"learning_rate": 0.0002131552646023783,
|
| 1123 |
+
"loss": 0.3391,
|
| 1124 |
+
"step": 1860
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.13,
|
| 1128 |
+
"learning_rate": 0.00021116430099863277,
|
| 1129 |
+
"loss": 0.3358,
|
| 1130 |
+
"step": 1870
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.13,
|
| 1134 |
+
"learning_rate": 0.0002091722269255404,
|
| 1135 |
+
"loss": 0.3364,
|
| 1136 |
+
"step": 1880
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.13,
|
| 1140 |
+
"learning_rate": 0.0002071792405269165,
|
| 1141 |
+
"loss": 0.341,
|
| 1142 |
+
"step": 1890
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.13,
|
| 1146 |
+
"learning_rate": 0.00020518554003732167,
|
| 1147 |
+
"loss": 0.3378,
|
| 1148 |
+
"step": 1900
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.13,
|
| 1152 |
+
"learning_rate": 0.00020319132376234462,
|
| 1153 |
+
"loss": 0.3305,
|
| 1154 |
+
"step": 1910
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.13,
|
| 1158 |
+
"learning_rate": 0.00020119679005887702,
|
| 1159 |
+
"loss": 0.3352,
|
| 1160 |
+
"step": 1920
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.14,
|
| 1164 |
+
"learning_rate": 0.00019920213731538394,
|
| 1165 |
+
"loss": 0.3376,
|
| 1166 |
+
"step": 1930
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.14,
|
| 1170 |
+
"learning_rate": 0.00019720756393217098,
|
| 1171 |
+
"loss": 0.3283,
|
| 1172 |
+
"step": 1940
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.14,
|
| 1176 |
+
"learning_rate": 0.00019521326830164998,
|
| 1177 |
+
"loss": 0.3354,
|
| 1178 |
+
"step": 1950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.14,
|
| 1182 |
+
"learning_rate": 0.00019321944878860587,
|
| 1183 |
+
"loss": 0.3445,
|
| 1184 |
+
"step": 1960
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.14,
|
| 1188 |
+
"learning_rate": 0.000191226303710466,
|
| 1189 |
+
"loss": 0.3401,
|
| 1190 |
+
"step": 1970
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.14,
|
| 1194 |
+
"learning_rate": 0.00018923403131757439,
|
| 1195 |
+
"loss": 0.3438,
|
| 1196 |
+
"step": 1980
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.14,
|
| 1200 |
+
"learning_rate": 0.00018724282977347235,
|
| 1201 |
+
"loss": 0.3341,
|
| 1202 |
+
"step": 1990
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.14,
|
| 1206 |
+
"learning_rate": 0.00018525289713518817,
|
| 1207 |
+
"loss": 0.3272,
|
| 1208 |
+
"step": 2000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.14,
|
| 1212 |
+
"learning_rate": 0.00018326443133353693,
|
| 1213 |
+
"loss": 0.3488,
|
| 1214 |
+
"step": 2010
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.14,
|
| 1218 |
+
"learning_rate": 0.00018127763015343332,
|
| 1219 |
+
"loss": 0.3242,
|
| 1220 |
+
"step": 2020
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.14,
|
| 1224 |
+
"learning_rate": 0.00017929269121421857,
|
| 1225 |
+
"loss": 0.337,
|
| 1226 |
+
"step": 2030
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.14,
|
| 1230 |
+
"learning_rate": 0.00017730981195000406,
|
| 1231 |
+
"loss": 0.3386,
|
| 1232 |
+
"step": 2040
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.14,
|
| 1236 |
+
"learning_rate": 0.00017532918959003353,
|
| 1237 |
+
"loss": 0.3312,
|
| 1238 |
+
"step": 2050
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.14,
|
| 1242 |
+
"learning_rate": 0.00017335102113906505,
|
| 1243 |
+
"loss": 0.3382,
|
| 1244 |
+
"step": 2060
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.15,
|
| 1248 |
+
"learning_rate": 0.00017137550335777612,
|
| 1249 |
+
"loss": 0.3319,
|
| 1250 |
+
"step": 2070
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.15,
|
| 1254 |
+
"learning_rate": 0.0001694028327431924,
|
| 1255 |
+
"loss": 0.3294,
|
| 1256 |
+
"step": 2080
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.15,
|
| 1260 |
+
"learning_rate": 0.0001674332055091431,
|
| 1261 |
+
"loss": 0.3313,
|
| 1262 |
+
"step": 2090
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.15,
|
| 1266 |
+
"learning_rate": 0.0001654668175667442,
|
| 1267 |
+
"loss": 0.3347,
|
| 1268 |
+
"step": 2100
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.15,
|
| 1272 |
+
"learning_rate": 0.00016350386450491208,
|
| 1273 |
+
"loss": 0.324,
|
| 1274 |
+
"step": 2110
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.15,
|
| 1278 |
+
"learning_rate": 0.00016154454157090884,
|
| 1279 |
+
"loss": 0.3302,
|
| 1280 |
+
"step": 2120
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.15,
|
| 1284 |
+
"learning_rate": 0.00015958904365092225,
|
| 1285 |
+
"loss": 0.334,
|
| 1286 |
+
"step": 2130
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.15,
|
| 1290 |
+
"learning_rate": 0.00015763756525068065,
|
| 1291 |
+
"loss": 0.3375,
|
| 1292 |
+
"step": 2140
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.15,
|
| 1296 |
+
"learning_rate": 0.00015569030047610656,
|
| 1297 |
+
"loss": 0.337,
|
| 1298 |
+
"step": 2150
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.15,
|
| 1302 |
+
"learning_rate": 0.0001537474430140096,
|
| 1303 |
+
"loss": 0.3346,
|
| 1304 |
+
"step": 2160
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.15,
|
| 1308 |
+
"learning_rate": 0.0001518091861128213,
|
| 1309 |
+
"loss": 0.3377,
|
| 1310 |
+
"step": 2170
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.15,
|
| 1314 |
+
"learning_rate": 0.00014987572256337336,
|
| 1315 |
+
"loss": 0.3247,
|
| 1316 |
+
"step": 2180
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.15,
|
| 1320 |
+
"learning_rate": 0.0001479472446797216,
|
| 1321 |
+
"loss": 0.3194,
|
| 1322 |
+
"step": 2190
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.15,
|
| 1326 |
+
"learning_rate": 0.00014602394428001712,
|
| 1327 |
+
"loss": 0.3299,
|
| 1328 |
+
"step": 2200
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.16,
|
| 1332 |
+
"learning_rate": 0.00014410601266742691,
|
| 1333 |
+
"loss": 0.3202,
|
| 1334 |
+
"step": 2210
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.16,
|
| 1338 |
+
"learning_rate": 0.00014219364061110565,
|
| 1339 |
+
"loss": 0.3316,
|
| 1340 |
+
"step": 2220
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.16,
|
| 1344 |
+
"learning_rate": 0.00014028701832722104,
|
| 1345 |
+
"loss": 0.3304,
|
| 1346 |
+
"step": 2230
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.16,
|
| 1350 |
+
"learning_rate": 0.00013838633546003302,
|
| 1351 |
+
"loss": 0.3286,
|
| 1352 |
+
"step": 2240
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.16,
|
| 1356 |
+
"learning_rate": 0.00013649178106303115,
|
| 1357 |
+
"loss": 0.3267,
|
| 1358 |
+
"step": 2250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.16,
|
| 1362 |
+
"learning_rate": 0.00013460354358013,
|
| 1363 |
+
"loss": 0.3295,
|
| 1364 |
+
"step": 2260
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.16,
|
| 1368 |
+
"learning_rate": 0.0001327218108269255,
|
| 1369 |
+
"loss": 0.3321,
|
| 1370 |
+
"step": 2270
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.16,
|
| 1374 |
+
"learning_rate": 0.00013084676997201342,
|
| 1375 |
+
"loss": 0.3302,
|
| 1376 |
+
"step": 2280
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.16,
|
| 1380 |
+
"learning_rate": 0.00012897860751837263,
|
| 1381 |
+
"loss": 0.3246,
|
| 1382 |
+
"step": 2290
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.16,
|
| 1386 |
+
"learning_rate": 0.00012711750928481443,
|
| 1387 |
+
"loss": 0.3305,
|
| 1388 |
+
"step": 2300
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.16,
|
| 1392 |
+
"learning_rate": 0.00012526366038749956,
|
| 1393 |
+
"loss": 0.3274,
|
| 1394 |
+
"step": 2310
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 0.16,
|
| 1398 |
+
"learning_rate": 0.000123417245221526,
|
| 1399 |
+
"loss": 0.3303,
|
| 1400 |
+
"step": 2320
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 0.16,
|
| 1404 |
+
"learning_rate": 0.00012157844744258722,
|
| 1405 |
+
"loss": 0.3265,
|
| 1406 |
+
"step": 2330
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.16,
|
| 1410 |
+
"learning_rate": 0.00011974744994870517,
|
| 1411 |
+
"loss": 0.3237,
|
| 1412 |
+
"step": 2340
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.16,
|
| 1416 |
+
"learning_rate": 0.00011792443486203788,
|
| 1417 |
+
"loss": 0.3288,
|
| 1418 |
+
"step": 2350
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.17,
|
| 1422 |
+
"learning_rate": 0.00011610958351076458,
|
| 1423 |
+
"loss": 0.3372,
|
| 1424 |
+
"step": 2360
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.17,
|
| 1428 |
+
"learning_rate": 0.00011430307641104971,
|
| 1429 |
+
"loss": 0.3257,
|
| 1430 |
+
"step": 2370
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.17,
|
| 1434 |
+
"learning_rate": 0.00011250509324908767,
|
| 1435 |
+
"loss": 0.3294,
|
| 1436 |
+
"step": 2380
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 0.17,
|
| 1440 |
+
"learning_rate": 0.00011071581286323,
|
| 1441 |
+
"loss": 0.3304,
|
| 1442 |
+
"step": 2390
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.17,
|
| 1446 |
+
"learning_rate": 0.00010893541322619732,
|
| 1447 |
+
"loss": 0.33,
|
| 1448 |
+
"step": 2400
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.17,
|
| 1452 |
+
"learning_rate": 0.00010716407142737659,
|
| 1453 |
+
"loss": 0.3251,
|
| 1454 |
+
"step": 2410
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.17,
|
| 1458 |
+
"learning_rate": 0.00010540196365520754,
|
| 1459 |
+
"loss": 0.3293,
|
| 1460 |
+
"step": 2420
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.17,
|
| 1464 |
+
"learning_rate": 0.00010364926517965692,
|
| 1465 |
+
"loss": 0.3269,
|
| 1466 |
+
"step": 2430
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.17,
|
| 1470 |
+
"learning_rate": 0.0001019061503347858,
|
| 1471 |
+
"loss": 0.3187,
|
| 1472 |
+
"step": 2440
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.17,
|
| 1476 |
+
"learning_rate": 0.00010017279250140891,
|
| 1477 |
+
"loss": 0.3204,
|
| 1478 |
+
"step": 2450
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 0.17,
|
| 1482 |
+
"learning_rate": 9.844936408984924e-05,
|
| 1483 |
+
"loss": 0.3151,
|
| 1484 |
+
"step": 2460
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.17,
|
| 1488 |
+
"learning_rate": 9.673603652278904e-05,
|
| 1489 |
+
"loss": 0.3237,
|
| 1490 |
+
"step": 2470
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.17,
|
| 1494 |
+
"learning_rate": 9.503298021821905e-05,
|
| 1495 |
+
"loss": 0.3224,
|
| 1496 |
+
"step": 2480
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.17,
|
| 1500 |
+
"learning_rate": 9.334036457248774e-05,
|
| 1501 |
+
"loss": 0.3228,
|
| 1502 |
+
"step": 2490
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.18,
|
| 1506 |
+
"learning_rate": 9.165835794345205e-05,
|
| 1507 |
+
"loss": 0.3186,
|
| 1508 |
+
"step": 2500
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.18,
|
| 1512 |
+
"learning_rate": 8.99871276337315e-05,
|
| 1513 |
+
"loss": 0.3181,
|
| 1514 |
+
"step": 2510
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.18,
|
| 1518 |
+
"learning_rate": 8.832683987406746e-05,
|
| 1519 |
+
"loss": 0.3233,
|
| 1520 |
+
"step": 2520
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.18,
|
| 1524 |
+
"learning_rate": 8.667765980678851e-05,
|
| 1525 |
+
"loss": 0.3302,
|
| 1526 |
+
"step": 2530
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.18,
|
| 1530 |
+
"learning_rate": 8.503975146938444e-05,
|
| 1531 |
+
"loss": 0.3314,
|
| 1532 |
+
"step": 2540
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.18,
|
| 1536 |
+
"learning_rate": 8.341327777819035e-05,
|
| 1537 |
+
"loss": 0.3309,
|
| 1538 |
+
"step": 2550
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.18,
|
| 1542 |
+
"learning_rate": 8.179840051218167e-05,
|
| 1543 |
+
"loss": 0.331,
|
| 1544 |
+
"step": 2560
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.18,
|
| 1548 |
+
"learning_rate": 8.019528029688286e-05,
|
| 1549 |
+
"loss": 0.318,
|
| 1550 |
+
"step": 2570
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.18,
|
| 1554 |
+
"learning_rate": 7.860407658839049e-05,
|
| 1555 |
+
"loss": 0.3273,
|
| 1556 |
+
"step": 2580
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.18,
|
| 1560 |
+
"learning_rate": 7.702494765751285e-05,
|
| 1561 |
+
"loss": 0.3182,
|
| 1562 |
+
"step": 2590
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.18,
|
| 1566 |
+
"learning_rate": 7.545805057402733e-05,
|
| 1567 |
+
"loss": 0.3237,
|
| 1568 |
+
"step": 2600
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.18,
|
| 1572 |
+
"learning_rate": 7.390354119105722e-05,
|
| 1573 |
+
"loss": 0.3229,
|
| 1574 |
+
"step": 2610
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.18,
|
| 1578 |
+
"learning_rate": 7.236157412956994e-05,
|
| 1579 |
+
"loss": 0.3304,
|
| 1580 |
+
"step": 2620
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.18,
|
| 1584 |
+
"learning_rate": 7.0832302762997e-05,
|
| 1585 |
+
"loss": 0.3284,
|
| 1586 |
+
"step": 2630
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.19,
|
| 1590 |
+
"learning_rate": 6.93158792019789e-05,
|
| 1591 |
+
"loss": 0.3251,
|
| 1592 |
+
"step": 2640
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.19,
|
| 1596 |
+
"learning_rate": 6.781245427923522e-05,
|
| 1597 |
+
"loss": 0.3201,
|
| 1598 |
+
"step": 2650
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.19,
|
| 1602 |
+
"learning_rate": 6.632217753456174e-05,
|
| 1603 |
+
"loss": 0.322,
|
| 1604 |
+
"step": 2660
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.19,
|
| 1608 |
+
"learning_rate": 6.484519719995647e-05,
|
| 1609 |
+
"loss": 0.3265,
|
| 1610 |
+
"step": 2670
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.19,
|
| 1614 |
+
"learning_rate": 6.338166018487555e-05,
|
| 1615 |
+
"loss": 0.3273,
|
| 1616 |
+
"step": 2680
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.19,
|
| 1620 |
+
"learning_rate": 6.193171206162065e-05,
|
| 1621 |
+
"loss": 0.3217,
|
| 1622 |
+
"step": 2690
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.19,
|
| 1626 |
+
"learning_rate": 6.0495497050859574e-05,
|
| 1627 |
+
"loss": 0.3247,
|
| 1628 |
+
"step": 2700
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.19,
|
| 1632 |
+
"learning_rate": 5.907315800728106e-05,
|
| 1633 |
+
"loss": 0.3179,
|
| 1634 |
+
"step": 2710
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.19,
|
| 1638 |
+
"learning_rate": 5.766483640538587e-05,
|
| 1639 |
+
"loss": 0.3116,
|
| 1640 |
+
"step": 2720
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.19,
|
| 1644 |
+
"learning_rate": 5.62706723254145e-05,
|
| 1645 |
+
"loss": 0.315,
|
| 1646 |
+
"step": 2730
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.19,
|
| 1650 |
+
"learning_rate": 5.489080443941415e-05,
|
| 1651 |
+
"loss": 0.3197,
|
| 1652 |
+
"step": 2740
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.19,
|
| 1656 |
+
"learning_rate": 5.352536999744557e-05,
|
| 1657 |
+
"loss": 0.32,
|
| 1658 |
+
"step": 2750
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 0.19,
|
| 1662 |
+
"learning_rate": 5.217450481393129e-05,
|
| 1663 |
+
"loss": 0.3195,
|
| 1664 |
+
"step": 2760
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.19,
|
| 1668 |
+
"learning_rate": 5.083834325414667e-05,
|
| 1669 |
+
"loss": 0.3221,
|
| 1670 |
+
"step": 2770
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.2,
|
| 1674 |
+
"learning_rate": 4.951701822085515e-05,
|
| 1675 |
+
"loss": 0.3165,
|
| 1676 |
+
"step": 2780
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.2,
|
| 1680 |
+
"learning_rate": 4.821066114108892e-05,
|
| 1681 |
+
"loss": 0.3216,
|
| 1682 |
+
"step": 2790
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.2,
|
| 1686 |
+
"learning_rate": 4.69194019530764e-05,
|
| 1687 |
+
"loss": 0.3122,
|
| 1688 |
+
"step": 2800
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.2,
|
| 1692 |
+
"learning_rate": 4.564336909331768e-05,
|
| 1693 |
+
"loss": 0.3237,
|
| 1694 |
+
"step": 2810
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.2,
|
| 1698 |
+
"learning_rate": 4.438268948380972e-05,
|
| 1699 |
+
"loss": 0.3169,
|
| 1700 |
+
"step": 2820
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 0.2,
|
| 1704 |
+
"learning_rate": 4.3137488519421656e-05,
|
| 1705 |
+
"loss": 0.3146,
|
| 1706 |
+
"step": 2830
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.2,
|
| 1710 |
+
"learning_rate": 4.1907890055422286e-05,
|
| 1711 |
+
"loss": 0.3256,
|
| 1712 |
+
"step": 2840
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.2,
|
| 1716 |
+
"learning_rate": 4.069401639516075e-05,
|
| 1717 |
+
"loss": 0.3178,
|
| 1718 |
+
"step": 2850
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.2,
|
| 1722 |
+
"learning_rate": 3.949598827790155e-05,
|
| 1723 |
+
"loss": 0.3151,
|
| 1724 |
+
"step": 2860
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.2,
|
| 1728 |
+
"learning_rate": 3.831392486681495e-05,
|
| 1729 |
+
"loss": 0.3121,
|
| 1730 |
+
"step": 2870
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.2,
|
| 1734 |
+
"learning_rate": 3.714794373712431e-05,
|
| 1735 |
+
"loss": 0.3199,
|
| 1736 |
+
"step": 2880
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.2,
|
| 1740 |
+
"learning_rate": 3.5998160864411476e-05,
|
| 1741 |
+
"loss": 0.3226,
|
| 1742 |
+
"step": 2890
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 0.2,
|
| 1746 |
+
"learning_rate": 3.486469061308093e-05,
|
| 1747 |
+
"loss": 0.3182,
|
| 1748 |
+
"step": 2900
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.2,
|
| 1752 |
+
"learning_rate": 3.3747645724984544e-05,
|
| 1753 |
+
"loss": 0.3162,
|
| 1754 |
+
"step": 2910
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.2,
|
| 1758 |
+
"learning_rate": 3.264713730820768e-05,
|
| 1759 |
+
"loss": 0.3133,
|
| 1760 |
+
"step": 2920
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.21,
|
| 1764 |
+
"learning_rate": 3.156327482601742e-05,
|
| 1765 |
+
"loss": 0.318,
|
| 1766 |
+
"step": 2930
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.21,
|
| 1770 |
+
"learning_rate": 3.0496166085974943e-05,
|
| 1771 |
+
"loss": 0.3107,
|
| 1772 |
+
"step": 2940
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.21,
|
| 1776 |
+
"learning_rate": 2.9445917229212193e-05,
|
| 1777 |
+
"loss": 0.3188,
|
| 1778 |
+
"step": 2950
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.21,
|
| 1782 |
+
"learning_rate": 2.8412632719874532e-05,
|
| 1783 |
+
"loss": 0.3278,
|
| 1784 |
+
"step": 2960
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.21,
|
| 1788 |
+
"learning_rate": 2.7396415334729964e-05,
|
| 1789 |
+
"loss": 0.3161,
|
| 1790 |
+
"step": 2970
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.21,
|
| 1794 |
+
"learning_rate": 2.6397366152946523e-05,
|
| 1795 |
+
"loss": 0.3151,
|
| 1796 |
+
"step": 2980
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.21,
|
| 1800 |
+
"learning_rate": 2.5415584546038096e-05,
|
| 1801 |
+
"loss": 0.316,
|
| 1802 |
+
"step": 2990
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.21,
|
| 1806 |
+
"learning_rate": 2.4451168167980497e-05,
|
| 1807 |
+
"loss": 0.3191,
|
| 1808 |
+
"step": 3000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.21,
|
| 1812 |
+
"learning_rate": 2.350421294549825e-05,
|
| 1813 |
+
"loss": 0.3152,
|
| 1814 |
+
"step": 3010
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 0.21,
|
| 1818 |
+
"learning_rate": 2.2574813068522894e-05,
|
| 1819 |
+
"loss": 0.3195,
|
| 1820 |
+
"step": 3020
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 0.21,
|
| 1824 |
+
"learning_rate": 2.166306098082451e-05,
|
| 1825 |
+
"loss": 0.3139,
|
| 1826 |
+
"step": 3030
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 0.21,
|
| 1830 |
+
"learning_rate": 2.0769047370816553e-05,
|
| 1831 |
+
"loss": 0.3214,
|
| 1832 |
+
"step": 3040
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.21,
|
| 1836 |
+
"learning_rate": 1.989286116253557e-05,
|
| 1837 |
+
"loss": 0.3114,
|
| 1838 |
+
"step": 3050
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.21,
|
| 1842 |
+
"learning_rate": 1.903458950679613e-05,
|
| 1843 |
+
"loss": 0.3203,
|
| 1844 |
+
"step": 3060
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.22,
|
| 1848 |
+
"learning_rate": 1.8194317772522362e-05,
|
| 1849 |
+
"loss": 0.3132,
|
| 1850 |
+
"step": 3070
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.22,
|
| 1854 |
+
"learning_rate": 1.7372129538256667e-05,
|
| 1855 |
+
"loss": 0.3091,
|
| 1856 |
+
"step": 3080
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 0.22,
|
| 1860 |
+
"learning_rate": 1.6568106583846378e-05,
|
| 1861 |
+
"loss": 0.3111,
|
| 1862 |
+
"step": 3090
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 0.22,
|
| 1866 |
+
"learning_rate": 1.5782328882309484e-05,
|
| 1867 |
+
"loss": 0.3126,
|
| 1868 |
+
"step": 3100
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 0.22,
|
| 1872 |
+
"learning_rate": 1.5014874591880157e-05,
|
| 1873 |
+
"loss": 0.3173,
|
| 1874 |
+
"step": 3110
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.22,
|
| 1878 |
+
"learning_rate": 1.4265820048234447e-05,
|
| 1879 |
+
"loss": 0.3084,
|
| 1880 |
+
"step": 3120
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.22,
|
| 1884 |
+
"learning_rate": 1.3535239756897566e-05,
|
| 1885 |
+
"loss": 0.3133,
|
| 1886 |
+
"step": 3130
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.22,
|
| 1890 |
+
"learning_rate": 1.2823206385833187e-05,
|
| 1891 |
+
"loss": 0.3182,
|
| 1892 |
+
"step": 3140
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.22,
|
| 1896 |
+
"learning_rate": 1.212979075821532e-05,
|
| 1897 |
+
"loss": 0.3109,
|
| 1898 |
+
"step": 3150
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 0.22,
|
| 1902 |
+
"learning_rate": 1.1455061845383852e-05,
|
| 1903 |
+
"loss": 0.3169,
|
| 1904 |
+
"step": 3160
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.22,
|
| 1908 |
+
"learning_rate": 1.0799086759984333e-05,
|
| 1909 |
+
"loss": 0.3185,
|
| 1910 |
+
"step": 3170
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 0.22,
|
| 1914 |
+
"learning_rate": 1.016193074929237e-05,
|
| 1915 |
+
"loss": 0.3035,
|
| 1916 |
+
"step": 3180
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.22,
|
| 1920 |
+
"learning_rate": 9.543657188723876e-06,
|
| 1921 |
+
"loss": 0.3117,
|
| 1922 |
+
"step": 3190
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.22,
|
| 1926 |
+
"learning_rate": 8.944327575531275e-06,
|
| 1927 |
+
"loss": 0.3204,
|
| 1928 |
+
"step": 3200
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.23,
|
| 1932 |
+
"learning_rate": 8.364001522686726e-06,
|
| 1933 |
+
"loss": 0.3144,
|
| 1934 |
+
"step": 3210
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.23,
|
| 1938 |
+
"learning_rate": 7.802736752952533e-06,
|
| 1939 |
+
"loss": 0.321,
|
| 1940 |
+
"step": 3220
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 0.23,
|
| 1944 |
+
"learning_rate": 7.260589093139736e-06,
|
| 1945 |
+
"loss": 0.316,
|
| 1946 |
+
"step": 3230
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 0.23,
|
| 1950 |
+
"learning_rate": 6.737612468555221e-06,
|
| 1951 |
+
"loss": 0.3133,
|
| 1952 |
+
"step": 3240
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 0.23,
|
| 1956 |
+
"learning_rate": 6.2338588976380115e-06,
|
| 1957 |
+
"loss": 0.3107,
|
| 1958 |
+
"step": 3250
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.23,
|
| 1962 |
+
"learning_rate": 5.749378486785162e-06,
|
| 1963 |
+
"loss": 0.3076,
|
| 1964 |
+
"step": 3260
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.23,
|
| 1968 |
+
"learning_rate": 5.2842194253679424e-06,
|
| 1969 |
+
"loss": 0.3137,
|
| 1970 |
+
"step": 3270
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.23,
|
| 1974 |
+
"learning_rate": 4.8384279809385426e-06,
|
| 1975 |
+
"loss": 0.3121,
|
| 1976 |
+
"step": 3280
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.23,
|
| 1980 |
+
"learning_rate": 4.41204849462804e-06,
|
| 1981 |
+
"loss": 0.3065,
|
| 1982 |
+
"step": 3290
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 0.23,
|
| 1986 |
+
"learning_rate": 4.005123376735997e-06,
|
| 1987 |
+
"loss": 0.3172,
|
| 1988 |
+
"step": 3300
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 0.23,
|
| 1992 |
+
"learning_rate": 3.617693102512032e-06,
|
| 1993 |
+
"loss": 0.3183,
|
| 1994 |
+
"step": 3310
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 0.23,
|
| 1998 |
+
"learning_rate": 3.2497962081299514e-06,
|
| 1999 |
+
"loss": 0.3141,
|
| 2000 |
+
"step": 3320
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.23,
|
| 2004 |
+
"learning_rate": 2.9014692868546633e-06,
|
| 2005 |
+
"loss": 0.3094,
|
| 2006 |
+
"step": 3330
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.23,
|
| 2010 |
+
"learning_rate": 2.572746985402419e-06,
|
| 2011 |
+
"loss": 0.3139,
|
| 2012 |
+
"step": 3340
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.24,
|
| 2016 |
+
"learning_rate": 2.2636620004946154e-06,
|
| 2017 |
+
"loss": 0.3091,
|
| 2018 |
+
"step": 3350
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.24,
|
| 2022 |
+
"learning_rate": 1.974245075605574e-06,
|
| 2023 |
+
"loss": 0.3092,
|
| 2024 |
+
"step": 3360
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 0.24,
|
| 2028 |
+
"learning_rate": 1.7045249979046995e-06,
|
| 2029 |
+
"loss": 0.3108,
|
| 2030 |
+
"step": 3370
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.24,
|
| 2034 |
+
"learning_rate": 1.4545285953929677e-06,
|
| 2035 |
+
"loss": 0.3098,
|
| 2036 |
+
"step": 3380
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 0.24,
|
| 2040 |
+
"learning_rate": 1.224280734234573e-06,
|
| 2041 |
+
"loss": 0.3059,
|
| 2042 |
+
"step": 3390
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.24,
|
| 2046 |
+
"learning_rate": 1.013804316283573e-06,
|
| 2047 |
+
"loss": 0.3073,
|
| 2048 |
+
"step": 3400
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.24,
|
| 2052 |
+
"learning_rate": 8.231202768059332e-07,
|
| 2053 |
+
"loss": 0.306,
|
| 2054 |
+
"step": 3410
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.24,
|
| 2058 |
+
"learning_rate": 6.522475823970808e-07,
|
| 2059 |
+
"loss": 0.3071,
|
| 2060 |
+
"step": 3420
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.24,
|
| 2064 |
+
"learning_rate": 5.012032290955037e-07,
|
| 2065 |
+
"loss": 0.3033,
|
| 2066 |
+
"step": 3430
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 0.24,
|
| 2070 |
+
"learning_rate": 3.7000224069216883e-07,
|
| 2071 |
+
"loss": 0.3109,
|
| 2072 |
+
"step": 3440
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.24,
|
| 2076 |
+
"learning_rate": 2.586576672361396e-07,
|
| 2077 |
+
"loss": 0.3127,
|
| 2078 |
+
"step": 3450
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 0.24,
|
| 2082 |
+
"learning_rate": 1.671805837365703e-07,
|
| 2083 |
+
"loss": 0.312,
|
| 2084 |
+
"step": 3460
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 0.24,
|
| 2088 |
+
"learning_rate": 9.558008906112026e-08,
|
| 2089 |
+
"loss": 0.317,
|
| 2090 |
+
"step": 3470
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.24,
|
| 2094 |
+
"learning_rate": 4.386330503090008e-08,
|
| 2095 |
+
"loss": 0.3092,
|
| 2096 |
+
"step": 3480
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.24,
|
| 2100 |
+
"learning_rate": 1.2035375712105001e-08,
|
| 2101 |
+
"loss": 0.3142,
|
| 2102 |
+
"step": 3490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.25,
|
| 2106 |
+
"learning_rate": 9.946690433526584e-11,
|
| 2107 |
+
"loss": 0.3105,
|
| 2108 |
+
"step": 3500
|
| 2109 |
+
}
|
| 2110 |
+
],
|
| 2111 |
+
"max_steps": 3500,
|
| 2112 |
+
"num_train_epochs": 1,
|
| 2113 |
+
"total_flos": 1.12189130145792e+18,
|
| 2114 |
+
"trial_name": null,
|
| 2115 |
+
"trial_params": null
|
| 2116 |
+
}
|
checkpoint-3500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05810fbd5b1488dde911cedadf47bfaefff7f71e2f51b45161b8d6e5275d232a
|
| 3 |
+
size 4664
|
runs/Nov24_18-45-12_0cb8a69e401c/events.out.tfevents.1700851546.0cb8a69e401c.2133.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d4142ad69d4de1a839cf09cae89e5e3b6104edd9f84f15106f9dfeaeb65d1b2b
|
| 3 |
+
size 4683
|
runs/Nov24_18-46-44_0cb8a69e401c/events.out.tfevents.1700851610.0cb8a69e401c.2133.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7cab738d8530b2ff201d28bc0d2e3bcc6c33a83fc0b373d6cf07bf5b22ccd64
|
| 3 |
+
size 5607
|
runs/Nov24_18-59-08_0cb8a69e401c/events.out.tfevents.1700852369.0cb8a69e401c.31161.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d1de15fe1c79375213117b3750e44ab5832d552875f3b4aaca25939aecccf1e
|
| 3 |
+
size 5299
|
runs/Nov25_00-57-56_0cb8a69e401c/events.out.tfevents.1700873879.0cb8a69e401c.31161.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e7b8dd47845240c7878e3f57663162d2ce5ca1f177ab0aae1c1f5e0a5aba166
|
| 3 |
+
size 4683
|
runs/Nov25_01-00-09_0cb8a69e401c/events.out.tfevents.1700874012.0cb8a69e401c.31161.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee591d492d8c21bbad3d764ac71120230e42485c535daf56da079c44591d533f
|
| 3 |
+
size 59597
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "<unk>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": true,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"__type": "AddedToken",
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"clean_up_tokenization_spaces": false,
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"__type": "AddedToken",
|
| 13 |
+
"content": "</s>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false
|
| 18 |
+
},
|
| 19 |
+
"legacy": false,
|
| 20 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 21 |
+
"pad_token": null,
|
| 22 |
+
"sp_model_kwargs": {},
|
| 23 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 24 |
+
"unk_token": {
|
| 25 |
+
"__type": "AddedToken",
|
| 26 |
+
"content": "<unk>",
|
| 27 |
+
"lstrip": false,
|
| 28 |
+
"normalized": true,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"single_word": false
|
| 31 |
+
}
|
| 32 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05810fbd5b1488dde911cedadf47bfaefff7f71e2f51b45161b8d6e5275d232a
|
| 3 |
+
size 4664
|