Hugging Face's logo

royallab
/
LimaRP-ShareGPT-13b-qloras

PEFT
Not-For-All-Audiences
Model card Files Community
2
kingbri commited on
Commit
f5c3cd0
•
1 Parent(s): b32633c

Add prompt-a twochar

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. prompt-a/twochar/README.md +21 -0
  2. prompt-a/twochar/adapter_config.json +26 -0
  3. prompt-a/twochar/adapter_model.bin +3 -0
  4. prompt-a/twochar/added_tokens.json +3 -0
  5. prompt-a/twochar/checkpoint-100/README.md +21 -0
  6. prompt-a/twochar/checkpoint-100/adapter_config.json +26 -0
  7. prompt-a/twochar/checkpoint-100/adapter_model.bin +3 -0
  8. prompt-a/twochar/checkpoint-100/adapter_model/README.md +21 -0
  9. prompt-a/twochar/checkpoint-100/adapter_model/adapter_config.json +26 -0
  10. prompt-a/twochar/checkpoint-100/adapter_model/adapter_model.bin +3 -0
  11. prompt-a/twochar/checkpoint-100/optimizer.pt +3 -0
  12. prompt-a/twochar/checkpoint-100/rng_state_0.pth +3 -0
  13. prompt-a/twochar/checkpoint-100/rng_state_1.pth +3 -0
  14. prompt-a/twochar/checkpoint-100/rng_state_2.pth +3 -0
  15. prompt-a/twochar/checkpoint-100/rng_state_3.pth +3 -0
  16. prompt-a/twochar/checkpoint-100/rng_state_4.pth +3 -0
  17. prompt-a/twochar/checkpoint-100/rng_state_5.pth +3 -0
  18. prompt-a/twochar/checkpoint-100/rng_state_6.pth +3 -0
  19. prompt-a/twochar/checkpoint-100/rng_state_7.pth +3 -0
  20. prompt-a/twochar/checkpoint-100/scheduler.pt +3 -0
  21. prompt-a/twochar/checkpoint-100/trainer_state.json +659 -0
  22. prompt-a/twochar/checkpoint-100/training_args.bin +3 -0
  23. prompt-a/twochar/checkpoint-150/README.md +21 -0
  24. prompt-a/twochar/checkpoint-150/adapter_config.json +26 -0
  25. prompt-a/twochar/checkpoint-150/adapter_model.bin +3 -0
  26. prompt-a/twochar/checkpoint-150/adapter_model/README.md +21 -0
  27. prompt-a/twochar/checkpoint-150/adapter_model/adapter_config.json +26 -0
  28. prompt-a/twochar/checkpoint-150/adapter_model/adapter_model.bin +3 -0
  29. prompt-a/twochar/checkpoint-150/optimizer.pt +3 -0
  30. prompt-a/twochar/checkpoint-150/rng_state_0.pth +3 -0
  31. prompt-a/twochar/checkpoint-150/rng_state_1.pth +3 -0
  32. prompt-a/twochar/checkpoint-150/rng_state_2.pth +3 -0
  33. prompt-a/twochar/checkpoint-150/rng_state_3.pth +3 -0
  34. prompt-a/twochar/checkpoint-150/rng_state_4.pth +3 -0
  35. prompt-a/twochar/checkpoint-150/rng_state_5.pth +3 -0
  36. prompt-a/twochar/checkpoint-150/rng_state_6.pth +3 -0
  37. prompt-a/twochar/checkpoint-150/rng_state_7.pth +3 -0
  38. prompt-a/twochar/checkpoint-150/scheduler.pt +3 -0
  39. prompt-a/twochar/checkpoint-150/trainer_state.json +975 -0
  40. prompt-a/twochar/checkpoint-150/training_args.bin +3 -0
  41. prompt-a/twochar/checkpoint-200/README.md +21 -0
  42. prompt-a/twochar/checkpoint-200/adapter_config.json +26 -0
  43. prompt-a/twochar/checkpoint-200/adapter_model.bin +3 -0
  44. prompt-a/twochar/checkpoint-200/adapter_model/README.md +21 -0
  45. prompt-a/twochar/checkpoint-200/adapter_model/adapter_config.json +26 -0
  46. prompt-a/twochar/checkpoint-200/adapter_model/adapter_model.bin +3 -0
  47. prompt-a/twochar/checkpoint-200/optimizer.pt +3 -0
  48. prompt-a/twochar/checkpoint-200/rng_state_0.pth +3 -0
  49. prompt-a/twochar/checkpoint-200/rng_state_1.pth +3 -0
  50. prompt-a/twochar/checkpoint-200/rng_state_2.pth +3 -0
prompt-a/twochar/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3686a6854a70ed7424cb6434eaf2d2d208980a99034b0d2075128de53b3297a2
3
+ size 125374989
prompt-a/twochar/added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "[PAD]": 32000
3
+ }
prompt-a/twochar/checkpoint-100/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-100/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-100/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42f483a8475f9c9615a60f6b89a84b8920c7a840df45ceabed0a3effadffc0e0
3
+ size 125374989
prompt-a/twochar/checkpoint-100/adapter_model/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-100/adapter_model/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-100/adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42f483a8475f9c9615a60f6b89a84b8920c7a840df45ceabed0a3effadffc0e0
3
+ size 125374989
prompt-a/twochar/checkpoint-100/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20d2e41d040cc30ba3448d2ef29ddff84499af170be7c5a74b063a9a719180ad
3
+ size 250681597
prompt-a/twochar/checkpoint-100/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f4ac5c72b3a8d37d6f5f4e72f59a981105e0739be3b7567447236a07702ef98
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09a9e7711a12aff1cb01365952914663e874439fc80ca6174d7f1ed92dd1fd03
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e71e7c3877097d7a81872b94110dacc2bb9179e8811b27061b892815e8d9aa49
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:101962960d17908eccf4d72170394b7c436c70f3676372f4a98047973af8b216
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3419ae90d4ffed73c9b8458beb5e60e1e5a986b8daf2d55fd3289f908f2ef5c9
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:064ab007d23c2b923f5dce2324188582329cc05945abb2006e33f8fbb94f27e0
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a231d16fbc0b8c521a5b6329b1042c9237bbaaf2cfddf194cc2c875474f5d4a9
3
+ size 21687
prompt-a/twochar/checkpoint-100/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab804b08c8130b997416c537045d5e5f947a41304867b421f39443b0b80131fe
3
+ size 21687
prompt-a/twochar/checkpoint-100/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b066d96ed0b5d7ac788c3a38bca16c4cd9e92b1f62ff59691c8a4867593fd5cf
3
+ size 627
prompt-a/twochar/checkpoint-100/trainer_state.json ADDED
@@ -0,0 +1,659 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 20,
6
+ "global_step": 100,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "learning_rate": 6.5e-06,
14
+ "loss": 1.9562,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.02,
19
+ "learning_rate": 1.3e-05,
20
+ "loss": 2.0273,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.03,
25
+ "learning_rate": 1.9499999999999996e-05,
26
+ "loss": 2.2516,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.04,
31
+ "learning_rate": 2.6e-05,
32
+ "loss": 2.1955,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.05,
37
+ "learning_rate": 3.25e-05,
38
+ "loss": 2.1752,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.06,
43
+ "learning_rate": 3.899999999999999e-05,
44
+ "loss": 2.2144,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.07,
49
+ "learning_rate": 4.5499999999999995e-05,
50
+ "loss": 2.0829,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.08,
55
+ "learning_rate": 5.2e-05,
56
+ "loss": 2.0503,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.09,
61
+ "learning_rate": 5.85e-05,
62
+ "loss": 2.1558,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.1,
67
+ "learning_rate": 6.5e-05,
68
+ "loss": 2.0924,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.11,
73
+ "learning_rate": 6.499555741225524e-05,
74
+ "loss": 2.1238,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.12,
79
+ "learning_rate": 6.498223086358009e-05,
80
+ "loss": 2.129,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.13,
85
+ "learning_rate": 6.49600239973199e-05,
86
+ "loss": 2.0559,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.14,
91
+ "learning_rate": 6.492894288461016e-05,
92
+ "loss": 2.0659,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.15,
97
+ "learning_rate": 6.488899602271677e-05,
98
+ "loss": 2.1563,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.16,
103
+ "learning_rate": 6.484019433271287e-05,
104
+ "loss": 2.1706,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.17,
109
+ "learning_rate": 6.478255115649324e-05,
110
+ "loss": 2.0736,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.18,
115
+ "learning_rate": 6.471608225312668e-05,
116
+ "loss": 2.1295,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.19,
121
+ "learning_rate": 6.464080579454772e-05,
122
+ "loss": 2.0618,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.2,
127
+ "learning_rate": 6.455674236058847e-05,
128
+ "loss": 2.12,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.2,
133
+ "eval_loss": 2.1955978870391846,
134
+ "eval_runtime": 2.9627,
135
+ "eval_samples_per_second": 3.375,
136
+ "eval_steps_per_second": 0.675,
137
+ "step": 20
138
+ },
139
+ {
140
+ "epoch": 0.21,
141
+ "learning_rate": 6.446391493335242e-05,
142
+ "loss": 2.1576,
143
+ "step": 21
144
+ },
145
+ {
146
+ "epoch": 0.22,
147
+ "learning_rate": 6.43623488909313e-05,
148
+ "loss": 2.0446,
149
+ "step": 22
150
+ },
151
+ {
152
+ "epoch": 0.23,
153
+ "learning_rate": 6.425207200046697e-05,
154
+ "loss": 2.0033,
155
+ "step": 23
156
+ },
157
+ {
158
+ "epoch": 0.24,
159
+ "learning_rate": 6.413311441056019e-05,
160
+ "loss": 2.1382,
161
+ "step": 24
162
+ },
163
+ {
164
+ "epoch": 0.25,
165
+ "learning_rate": 6.400550864302823e-05,
166
+ "loss": 2.0989,
167
+ "step": 25
168
+ },
169
+ {
170
+ "epoch": 0.26,
171
+ "learning_rate": 6.386928958401383e-05,
172
+ "loss": 2.1276,
173
+ "step": 26
174
+ },
175
+ {
176
+ "epoch": 0.27,
177
+ "learning_rate": 6.372449447444757e-05,
178
+ "loss": 2.023,
179
+ "step": 27
180
+ },
181
+ {
182
+ "epoch": 0.28,
183
+ "learning_rate": 6.357116289986662e-05,
184
+ "loss": 2.0597,
185
+ "step": 28
186
+ },
187
+ {
188
+ "epoch": 0.29,
189
+ "learning_rate": 6.340933677959248e-05,
190
+ "loss": 2.0529,
191
+ "step": 29
192
+ },
193
+ {
194
+ "epoch": 0.3,
195
+ "learning_rate": 6.323906035527062e-05,
196
+ "loss": 1.9873,
197
+ "step": 30
198
+ },
199
+ {
200
+ "epoch": 0.31,
201
+ "learning_rate": 6.306038017877523e-05,
202
+ "loss": 2.0613,
203
+ "step": 31
204
+ },
205
+ {
206
+ "epoch": 0.32,
207
+ "learning_rate": 6.287334509948247e-05,
208
+ "loss": 2.0765,
209
+ "step": 32
210
+ },
211
+ {
212
+ "epoch": 0.33,
213
+ "learning_rate": 6.267800625091547e-05,
214
+ "loss": 2.0809,
215
+ "step": 33
216
+ },
217
+ {
218
+ "epoch": 0.34,
219
+ "learning_rate": 6.2474417036765e-05,
220
+ "loss": 2.072,
221
+ "step": 34
222
+ },
223
+ {
224
+ "epoch": 0.35,
225
+ "learning_rate": 6.226263311628936e-05,
226
+ "loss": 2.0069,
227
+ "step": 35
228
+ },
229
+ {
230
+ "epoch": 0.36,
231
+ "learning_rate": 6.204271238909777e-05,
232
+ "loss": 2.0784,
233
+ "step": 36
234
+ },
235
+ {
236
+ "epoch": 0.37,
237
+ "learning_rate": 6.181471497932117e-05,
238
+ "loss": 2.1763,
239
+ "step": 37
240
+ },
241
+ {
242
+ "epoch": 0.38,
243
+ "learning_rate": 6.157870321917481e-05,
244
+ "loss": 2.0402,
245
+ "step": 38
246
+ },
247
+ {
248
+ "epoch": 0.39,
249
+ "learning_rate": 6.133474163191738e-05,
250
+ "loss": 2.1082,
251
+ "step": 39
252
+ },
253
+ {
254
+ "epoch": 0.4,
255
+ "learning_rate": 6.108289691421088e-05,
256
+ "loss": 2.1381,
257
+ "step": 40
258
+ },
259
+ {
260
+ "epoch": 0.4,
261
+ "eval_loss": 2.172107219696045,
262
+ "eval_runtime": 2.97,
263
+ "eval_samples_per_second": 3.367,
264
+ "eval_steps_per_second": 0.673,
265
+ "step": 40
266
+ },
267
+ {
268
+ "epoch": 0.41,
269
+ "learning_rate": 6.08232379178865e-05,
270
+ "loss": 2.0538,
271
+ "step": 41
272
+ },
273
+ {
274
+ "epoch": 0.42,
275
+ "learning_rate": 6.055583563112115e-05,
276
+ "loss": 2.222,
277
+ "step": 42
278
+ },
279
+ {
280
+ "epoch": 0.43,
281
+ "learning_rate": 6.0280763159030015e-05,
282
+ "loss": 1.996,
283
+ "step": 43
284
+ },
285
+ {
286
+ "epoch": 0.44,
287
+ "learning_rate": 5.999809570368039e-05,
288
+ "loss": 1.9798,
289
+ "step": 44
290
+ },
291
+ {
292
+ "epoch": 0.45,
293
+ "learning_rate": 5.970791054353217e-05,
294
+ "loss": 2.1825,
295
+ "step": 45
296
+ },
297
+ {
298
+ "epoch": 0.46,
299
+ "learning_rate": 5.941028701231065e-05,
300
+ "loss": 2.1671,
301
+ "step": 46
302
+ },
303
+ {
304
+ "epoch": 0.47,
305
+ "learning_rate": 5.91053064773175e-05,
306
+ "loss": 2.0992,
307
+ "step": 47
308
+ },
309
+ {
310
+ "epoch": 0.48,
311
+ "learning_rate": 5.8793052317185785e-05,
312
+ "loss": 2.0544,
313
+ "step": 48
314
+ },
315
+ {
316
+ "epoch": 0.49,
317
+ "learning_rate": 5.8473609899085046e-05,
318
+ "loss": 2.1054,
319
+ "step": 49
320
+ },
321
+ {
322
+ "epoch": 0.5,
323
+ "learning_rate": 5.814706655538279e-05,
324
+ "loss": 2.1214,
325
+ "step": 50
326
+ },
327
+ {
328
+ "epoch": 0.51,
329
+ "learning_rate": 5.781351155976865e-05,
330
+ "loss": 1.9961,
331
+ "step": 51
332
+ },
333
+ {
334
+ "epoch": 0.52,
335
+ "learning_rate": 5.7473036102847915e-05,
336
+ "loss": 2.0233,
337
+ "step": 52
338
+ },
339
+ {
340
+ "epoch": 0.53,
341
+ "learning_rate": 5.712573326721088e-05,
342
+ "loss": 2.0756,
343
+ "step": 53
344
+ },
345
+ {
346
+ "epoch": 0.54,
347
+ "learning_rate": 5.6771698001984996e-05,
348
+ "loss": 2.0138,
349
+ "step": 54
350
+ },
351
+ {
352
+ "epoch": 0.55,
353
+ "learning_rate": 5.641102709687678e-05,
354
+ "loss": 1.9964,
355
+ "step": 55
356
+ },
357
+ {
358
+ "epoch": 0.56,
359
+ "learning_rate": 5.604381915571039e-05,
360
+ "loss": 2.0915,
361
+ "step": 56
362
+ },
363
+ {
364
+ "epoch": 0.57,
365
+ "learning_rate": 5.5670174569470405e-05,
366
+ "loss": 1.9954,
367
+ "step": 57
368
+ },
369
+ {
370
+ "epoch": 0.58,
371
+ "learning_rate": 5.529019548885587e-05,
372
+ "loss": 2.1395,
373
+ "step": 58
374
+ },
375
+ {
376
+ "epoch": 0.59,
377
+ "learning_rate": 5.490398579635338e-05,
378
+ "loss": 2.0388,
379
+ "step": 59
380
+ },
381
+ {
382
+ "epoch": 0.6,
383
+ "learning_rate": 5.451165107783658e-05,
384
+ "loss": 2.0979,
385
+ "step": 60
386
+ },
387
+ {
388
+ "epoch": 0.6,
389
+ "eval_loss": 2.166550874710083,
390
+ "eval_runtime": 2.9777,
391
+ "eval_samples_per_second": 3.358,
392
+ "eval_steps_per_second": 0.672,
393
+ "step": 60
394
+ },
395
+ {
396
+ "epoch": 0.61,
397
+ "learning_rate": 5.411329859370006e-05,
398
+ "loss": 2.0614,
399
+ "step": 61
400
+ },
401
+ {
402
+ "epoch": 0.62,
403
+ "learning_rate": 5.370903724953548e-05,
404
+ "loss": 2.061,
405
+ "step": 62
406
+ },
407
+ {
408
+ "epoch": 0.63,
409
+ "learning_rate": 5.329897756635779e-05,
410
+ "loss": 2.0868,
411
+ "step": 63
412
+ },
413
+ {
414
+ "epoch": 0.64,
415
+ "learning_rate": 5.2883231650389946e-05,
416
+ "loss": 2.0659,
417
+ "step": 64
418
+ },
419
+ {
420
+ "epoch": 0.65,
421
+ "learning_rate": 5.2461913162414206e-05,
422
+ "loss": 2.1325,
423
+ "step": 65
424
+ },
425
+ {
426
+ "epoch": 0.66,
427
+ "learning_rate": 5.203513728669831e-05,
428
+ "loss": 2.0955,
429
+ "step": 66
430
+ },
431
+ {
432
+ "epoch": 0.67,
433
+ "learning_rate": 5.160302069950537e-05,
434
+ "loss": 1.9692,
435
+ "step": 67
436
+ },
437
+ {
438
+ "epoch": 0.68,
439
+ "learning_rate": 5.116568153719566e-05,
440
+ "loss": 2.0858,
441
+ "step": 68
442
+ },
443
+ {
444
+ "epoch": 0.69,
445
+ "learning_rate": 5.072323936392936e-05,
446
+ "loss": 2.0884,
447
+ "step": 69
448
+ },
449
+ {
450
+ "epoch": 0.7,
451
+ "learning_rate": 5.027581513897888e-05,
452
+ "loss": 1.952,
453
+ "step": 70
454
+ },
455
+ {
456
+ "epoch": 0.71,
457
+ "learning_rate": 4.9823531183659796e-05,
458
+ "loss": 2.0223,
459
+ "step": 71
460
+ },
461
+ {
462
+ "epoch": 0.72,
463
+ "learning_rate": 4.936651114788949e-05,
464
+ "loss": 2.0098,
465
+ "step": 72
466
+ },
467
+ {
468
+ "epoch": 0.73,
469
+ "learning_rate": 4.890487997638242e-05,
470
+ "loss": 1.9965,
471
+ "step": 73
472
+ },
473
+ {
474
+ "epoch": 0.74,
475
+ "learning_rate": 4.843876387449151e-05,
476
+ "loss": 2.1097,
477
+ "step": 74
478
+ },
479
+ {
480
+ "epoch": 0.75,
481
+ "learning_rate": 4.796829027370489e-05,
482
+ "loss": 2.1051,
483
+ "step": 75
484
+ },
485
+ {
486
+ "epoch": 0.76,
487
+ "learning_rate": 4.749358779680735e-05,
488
+ "loss": 2.1672,
489
+ "step": 76
490
+ },
491
+ {
492
+ "epoch": 0.77,
493
+ "learning_rate": 4.701478622271621e-05,
494
+ "loss": 2.0641,
495
+ "step": 77
496
+ },
497
+ {
498
+ "epoch": 0.78,
499
+ "learning_rate": 4.653201645100101e-05,
500
+ "loss": 2.1899,
501
+ "step": 78
502
+ },
503
+ {
504
+ "epoch": 0.79,
505
+ "learning_rate": 4.6045410466096934e-05,
506
+ "loss": 2.007,
507
+ "step": 79
508
+ },
509
+ {
510
+ "epoch": 0.8,
511
+ "learning_rate": 4.555510130122151e-05,
512
+ "loss": 2.1355,
513
+ "step": 80
514
+ },
515
+ {
516
+ "epoch": 0.8,
517
+ "eval_loss": 2.160327434539795,
518
+ "eval_runtime": 2.9723,
519
+ "eval_samples_per_second": 3.364,
520
+ "eval_steps_per_second": 0.673,
521
+ "step": 80
522
+ },
523
+ {
524
+ "epoch": 0.81,
525
+ "learning_rate": 4.5061223002004694e-05,
526
+ "loss": 2.0046,
527
+ "step": 81
528
+ },
529
+ {
530
+ "epoch": 0.82,
531
+ "learning_rate": 4.456391058984217e-05,
532
+ "loss": 2.0902,
533
+ "step": 82
534
+ },
535
+ {
536
+ "epoch": 0.83,
537
+ "learning_rate": 4.406330002498178e-05,
538
+ "loss": 2.1259,
539
+ "step": 83
540
+ },
541
+ {
542
+ "epoch": 0.84,
543
+ "learning_rate": 4.3559528169353394e-05,
544
+ "loss": 1.9241,
545
+ "step": 84
546
+ },
547
+ {
548
+ "epoch": 0.85,
549
+ "learning_rate": 4.3052732749152215e-05,
550
+ "loss": 2.1537,
551
+ "step": 85
552
+ },
553
+ {
554
+ "epoch": 0.86,
555
+ "learning_rate": 4.254305231718579e-05,
556
+ "loss": 1.9482,
557
+ "step": 86
558
+ },
559
+ {
560
+ "epoch": 0.87,
561
+ "learning_rate": 4.2030626214995104e-05,
562
+ "loss": 2.0181,
563
+ "step": 87
564
+ },
565
+ {
566
+ "epoch": 0.88,
567
+ "learning_rate": 4.151559453475994e-05,
568
+ "loss": 1.9839,
569
+ "step": 88
570
+ },
571
+ {
572
+ "epoch": 0.89,
573
+ "learning_rate": 4.0998098080999136e-05,
574
+ "loss": 2.0735,
575
+ "step": 89
576
+ },
577
+ {
578
+ "epoch": 0.9,
579
+ "learning_rate": 4.047827833207597e-05,
580
+ "loss": 2.1691,
581
+ "step": 90
582
+ },
583
+ {
584
+ "epoch": 0.91,
585
+ "learning_rate": 3.995627740151942e-05,
586
+ "loss": 2.2761,
587
+ "step": 91
588
+ },
589
+ {
590
+ "epoch": 0.92,
591
+ "learning_rate": 3.9432237999171774e-05,
592
+ "loss": 2.1657,
593
+ "step": 92
594
+ },
595
+ {
596
+ "epoch": 0.93,
597
+ "learning_rate": 3.89063033921731e-05,
598
+ "loss": 1.9501,
599
+ "step": 93
600
+ },
601
+ {
602
+ "epoch": 0.94,
603
+ "learning_rate": 3.8378617365793576e-05,
604
+ "loss": 2.0869,
605
+ "step": 94
606
+ },
607
+ {
608
+ "epoch": 0.95,
609
+ "learning_rate": 3.784932418412385e-05,
610
+ "loss": 2.0112,
611
+ "step": 95
612
+ },
613
+ {
614
+ "epoch": 0.96,
615
+ "learning_rate": 3.731856855063486e-05,
616
+ "loss": 1.9212,
617
+ "step": 96
618
+ },
619
+ {
620
+ "epoch": 0.97,
621
+ "learning_rate": 3.678649556861724e-05,
622
+ "loss": 2.0395,
623
+ "step": 97
624
+ },
625
+ {
626
+ "epoch": 0.98,
627
+ "learning_rate": 3.625325070151155e-05,
628
+ "loss": 2.0179,
629
+ "step": 98
630
+ },
631
+ {
632
+ "epoch": 0.99,
633
+ "learning_rate": 3.571897973314004e-05,
634
+ "loss": 1.9713,
635
+ "step": 99
636
+ },
637
+ {
638
+ "epoch": 1.0,
639
+ "learning_rate": 3.51838287278508e-05,
640
+ "loss": 2.0756,
641
+ "step": 100
642
+ },
643
+ {
644
+ "epoch": 1.0,
645
+ "eval_loss": 2.1577982902526855,
646
+ "eval_runtime": 2.9721,
647
+ "eval_samples_per_second": 3.365,
648
+ "eval_steps_per_second": 0.673,
649
+ "step": 100
650
+ }
651
+ ],
652
+ "logging_steps": 1,
653
+ "max_steps": 200,
654
+ "num_train_epochs": 2,
655
+ "save_steps": 50,
656
+ "total_flos": 8.361608023638016e+16,
657
+ "trial_name": null,
658
+ "trial_params": null
659
+ }
prompt-a/twochar/checkpoint-100/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9212c1ed1ebf30dd523c0f488bee8e1daf66e67eaaf76d60029260d57f02f88
3
+ size 4411
prompt-a/twochar/checkpoint-150/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-150/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-150/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef2308923705e0db072fa56c1601591d86bd453af7564a28a77b84dd1d79e8c4
3
+ size 125374989
prompt-a/twochar/checkpoint-150/adapter_model/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-150/adapter_model/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-150/adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef2308923705e0db072fa56c1601591d86bd453af7564a28a77b84dd1d79e8c4
3
+ size 125374989
prompt-a/twochar/checkpoint-150/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:602ac917af5d4639ca282f520dde56e364bf3f0db429735790992bf32ae66b1d
3
+ size 250681597
prompt-a/twochar/checkpoint-150/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89eadc3c165657fd8751b85f9ff44d7c95cc36af0669fbff4a4ffe5e0b07f5fc
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3745b667b2a6ca4a67753c6c91a46d6482d85c16a72b363be05fc59f7c442c14
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:730ab402a78e7c0ffa824fcd5f320e6ac0fca6070a3982c84cac9eb233003d70
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6ee13cc98624b772f8f4143d6dccdc5cdaff35faa989b04753b6d57e692fa4c
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c7a46010650ea94fbb45a5cb40a5d7a9cc9a325ad0dcbbe32f26384cfedc3c9
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:448b2bc0d10409f11bb41fd0844a35fce476a3ffaa24967c8dfa611c6f55547d
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2d6cd239229ac1dd36dd86fc7bed222b898f45d69914c4668c12378f553f4be
3
+ size 21687
prompt-a/twochar/checkpoint-150/rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1f9bd38de615aa1185c1abc88320d66fe7f03cb4281a215b715cb5b31205b2e
3
+ size 21687
prompt-a/twochar/checkpoint-150/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31da033238aea895275b0ce325aba2b20a03c1ced0b3d0cc962cb556ea08a8a9
3
+ size 627
prompt-a/twochar/checkpoint-150/trainer_state.json ADDED
@@ -0,0 +1,975 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.5,
5
+ "eval_steps": 20,
6
+ "global_step": 150,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "learning_rate": 6.5e-06,
14
+ "loss": 1.9562,
15
+ "step": 1
16
+ },
17
+ {
18
+ "epoch": 0.02,
19
+ "learning_rate": 1.3e-05,
20
+ "loss": 2.0273,
21
+ "step": 2
22
+ },
23
+ {
24
+ "epoch": 0.03,
25
+ "learning_rate": 1.9499999999999996e-05,
26
+ "loss": 2.2516,
27
+ "step": 3
28
+ },
29
+ {
30
+ "epoch": 0.04,
31
+ "learning_rate": 2.6e-05,
32
+ "loss": 2.1955,
33
+ "step": 4
34
+ },
35
+ {
36
+ "epoch": 0.05,
37
+ "learning_rate": 3.25e-05,
38
+ "loss": 2.1752,
39
+ "step": 5
40
+ },
41
+ {
42
+ "epoch": 0.06,
43
+ "learning_rate": 3.899999999999999e-05,
44
+ "loss": 2.2144,
45
+ "step": 6
46
+ },
47
+ {
48
+ "epoch": 0.07,
49
+ "learning_rate": 4.5499999999999995e-05,
50
+ "loss": 2.0829,
51
+ "step": 7
52
+ },
53
+ {
54
+ "epoch": 0.08,
55
+ "learning_rate": 5.2e-05,
56
+ "loss": 2.0503,
57
+ "step": 8
58
+ },
59
+ {
60
+ "epoch": 0.09,
61
+ "learning_rate": 5.85e-05,
62
+ "loss": 2.1558,
63
+ "step": 9
64
+ },
65
+ {
66
+ "epoch": 0.1,
67
+ "learning_rate": 6.5e-05,
68
+ "loss": 2.0924,
69
+ "step": 10
70
+ },
71
+ {
72
+ "epoch": 0.11,
73
+ "learning_rate": 6.499555741225524e-05,
74
+ "loss": 2.1238,
75
+ "step": 11
76
+ },
77
+ {
78
+ "epoch": 0.12,
79
+ "learning_rate": 6.498223086358009e-05,
80
+ "loss": 2.129,
81
+ "step": 12
82
+ },
83
+ {
84
+ "epoch": 0.13,
85
+ "learning_rate": 6.49600239973199e-05,
86
+ "loss": 2.0559,
87
+ "step": 13
88
+ },
89
+ {
90
+ "epoch": 0.14,
91
+ "learning_rate": 6.492894288461016e-05,
92
+ "loss": 2.0659,
93
+ "step": 14
94
+ },
95
+ {
96
+ "epoch": 0.15,
97
+ "learning_rate": 6.488899602271677e-05,
98
+ "loss": 2.1563,
99
+ "step": 15
100
+ },
101
+ {
102
+ "epoch": 0.16,
103
+ "learning_rate": 6.484019433271287e-05,
104
+ "loss": 2.1706,
105
+ "step": 16
106
+ },
107
+ {
108
+ "epoch": 0.17,
109
+ "learning_rate": 6.478255115649324e-05,
110
+ "loss": 2.0736,
111
+ "step": 17
112
+ },
113
+ {
114
+ "epoch": 0.18,
115
+ "learning_rate": 6.471608225312668e-05,
116
+ "loss": 2.1295,
117
+ "step": 18
118
+ },
119
+ {
120
+ "epoch": 0.19,
121
+ "learning_rate": 6.464080579454772e-05,
122
+ "loss": 2.0618,
123
+ "step": 19
124
+ },
125
+ {
126
+ "epoch": 0.2,
127
+ "learning_rate": 6.455674236058847e-05,
128
+ "loss": 2.12,
129
+ "step": 20
130
+ },
131
+ {
132
+ "epoch": 0.2,
133
+ "eval_loss": 2.1955978870391846,
134
+ "eval_runtime": 2.9627,
135
+ "eval_samples_per_second": 3.375,
136
+ "eval_steps_per_second": 0.675,
137
+ "step": 20
138
+ },
139
+ {
140
+ "epoch": 0.21,
141
+ "learning_rate": 6.446391493335242e-05,
142
+ "loss": 2.1576,
143
+ "step": 21
144
+ },
145
+ {
146
+ "epoch": 0.22,
147
+ "learning_rate": 6.43623488909313e-05,
148
+ "loss": 2.0446,
149
+ "step": 22
150
+ },
151
+ {
152
+ "epoch": 0.23,
153
+ "learning_rate": 6.425207200046697e-05,
154
+ "loss": 2.0033,
155
+ "step": 23
156
+ },
157
+ {
158
+ "epoch": 0.24,
159
+ "learning_rate": 6.413311441056019e-05,
160
+ "loss": 2.1382,
161
+ "step": 24
162
+ },
163
+ {
164
+ "epoch": 0.25,
165
+ "learning_rate": 6.400550864302823e-05,
166
+ "loss": 2.0989,
167
+ "step": 25
168
+ },
169
+ {
170
+ "epoch": 0.26,
171
+ "learning_rate": 6.386928958401383e-05,
172
+ "loss": 2.1276,
173
+ "step": 26
174
+ },
175
+ {
176
+ "epoch": 0.27,
177
+ "learning_rate": 6.372449447444757e-05,
178
+ "loss": 2.023,
179
+ "step": 27
180
+ },
181
+ {
182
+ "epoch": 0.28,
183
+ "learning_rate": 6.357116289986662e-05,
184
+ "loss": 2.0597,
185
+ "step": 28
186
+ },
187
+ {
188
+ "epoch": 0.29,
189
+ "learning_rate": 6.340933677959248e-05,
190
+ "loss": 2.0529,
191
+ "step": 29
192
+ },
193
+ {
194
+ "epoch": 0.3,
195
+ "learning_rate": 6.323906035527062e-05,
196
+ "loss": 1.9873,
197
+ "step": 30
198
+ },
199
+ {
200
+ "epoch": 0.31,
201
+ "learning_rate": 6.306038017877523e-05,
202
+ "loss": 2.0613,
203
+ "step": 31
204
+ },
205
+ {
206
+ "epoch": 0.32,
207
+ "learning_rate": 6.287334509948247e-05,
208
+ "loss": 2.0765,
209
+ "step": 32
210
+ },
211
+ {
212
+ "epoch": 0.33,
213
+ "learning_rate": 6.267800625091547e-05,
214
+ "loss": 2.0809,
215
+ "step": 33
216
+ },
217
+ {
218
+ "epoch": 0.34,
219
+ "learning_rate": 6.2474417036765e-05,
220
+ "loss": 2.072,
221
+ "step": 34
222
+ },
223
+ {
224
+ "epoch": 0.35,
225
+ "learning_rate": 6.226263311628936e-05,
226
+ "loss": 2.0069,
227
+ "step": 35
228
+ },
229
+ {
230
+ "epoch": 0.36,
231
+ "learning_rate": 6.204271238909777e-05,
232
+ "loss": 2.0784,
233
+ "step": 36
234
+ },
235
+ {
236
+ "epoch": 0.37,
237
+ "learning_rate": 6.181471497932117e-05,
238
+ "loss": 2.1763,
239
+ "step": 37
240
+ },
241
+ {
242
+ "epoch": 0.38,
243
+ "learning_rate": 6.157870321917481e-05,
244
+ "loss": 2.0402,
245
+ "step": 38
246
+ },
247
+ {
248
+ "epoch": 0.39,
249
+ "learning_rate": 6.133474163191738e-05,
250
+ "loss": 2.1082,
251
+ "step": 39
252
+ },
253
+ {
254
+ "epoch": 0.4,
255
+ "learning_rate": 6.108289691421088e-05,
256
+ "loss": 2.1381,
257
+ "step": 40
258
+ },
259
+ {
260
+ "epoch": 0.4,
261
+ "eval_loss": 2.172107219696045,
262
+ "eval_runtime": 2.97,
263
+ "eval_samples_per_second": 3.367,
264
+ "eval_steps_per_second": 0.673,
265
+ "step": 40
266
+ },
267
+ {
268
+ "epoch": 0.41,
269
+ "learning_rate": 6.08232379178865e-05,
270
+ "loss": 2.0538,
271
+ "step": 41
272
+ },
273
+ {
274
+ "epoch": 0.42,
275
+ "learning_rate": 6.055583563112115e-05,
276
+ "loss": 2.222,
277
+ "step": 42
278
+ },
279
+ {
280
+ "epoch": 0.43,
281
+ "learning_rate": 6.0280763159030015e-05,
282
+ "loss": 1.996,
283
+ "step": 43
284
+ },
285
+ {
286
+ "epoch": 0.44,
287
+ "learning_rate": 5.999809570368039e-05,
288
+ "loss": 1.9798,
289
+ "step": 44
290
+ },
291
+ {
292
+ "epoch": 0.45,
293
+ "learning_rate": 5.970791054353217e-05,
294
+ "loss": 2.1825,
295
+ "step": 45
296
+ },
297
+ {
298
+ "epoch": 0.46,
299
+ "learning_rate": 5.941028701231065e-05,
300
+ "loss": 2.1671,
301
+ "step": 46
302
+ },
303
+ {
304
+ "epoch": 0.47,
305
+ "learning_rate": 5.91053064773175e-05,
306
+ "loss": 2.0992,
307
+ "step": 47
308
+ },
309
+ {
310
+ "epoch": 0.48,
311
+ "learning_rate": 5.8793052317185785e-05,
312
+ "loss": 2.0544,
313
+ "step": 48
314
+ },
315
+ {
316
+ "epoch": 0.49,
317
+ "learning_rate": 5.8473609899085046e-05,
318
+ "loss": 2.1054,
319
+ "step": 49
320
+ },
321
+ {
322
+ "epoch": 0.5,
323
+ "learning_rate": 5.814706655538279e-05,
324
+ "loss": 2.1214,
325
+ "step": 50
326
+ },
327
+ {
328
+ "epoch": 0.51,
329
+ "learning_rate": 5.781351155976865e-05,
330
+ "loss": 1.9961,
331
+ "step": 51
332
+ },
333
+ {
334
+ "epoch": 0.52,
335
+ "learning_rate": 5.7473036102847915e-05,
336
+ "loss": 2.0233,
337
+ "step": 52
338
+ },
339
+ {
340
+ "epoch": 0.53,
341
+ "learning_rate": 5.712573326721088e-05,
342
+ "loss": 2.0756,
343
+ "step": 53
344
+ },
345
+ {
346
+ "epoch": 0.54,
347
+ "learning_rate": 5.6771698001984996e-05,
348
+ "loss": 2.0138,
349
+ "step": 54
350
+ },
351
+ {
352
+ "epoch": 0.55,
353
+ "learning_rate": 5.641102709687678e-05,
354
+ "loss": 1.9964,
355
+ "step": 55
356
+ },
357
+ {
358
+ "epoch": 0.56,
359
+ "learning_rate": 5.604381915571039e-05,
360
+ "loss": 2.0915,
361
+ "step": 56
362
+ },
363
+ {
364
+ "epoch": 0.57,
365
+ "learning_rate": 5.5670174569470405e-05,
366
+ "loss": 1.9954,
367
+ "step": 57
368
+ },
369
+ {
370
+ "epoch": 0.58,
371
+ "learning_rate": 5.529019548885587e-05,
372
+ "loss": 2.1395,
373
+ "step": 58
374
+ },
375
+ {
376
+ "epoch": 0.59,
377
+ "learning_rate": 5.490398579635338e-05,
378
+ "loss": 2.0388,
379
+ "step": 59
380
+ },
381
+ {
382
+ "epoch": 0.6,
383
+ "learning_rate": 5.451165107783658e-05,
384
+ "loss": 2.0979,
385
+ "step": 60
386
+ },
387
+ {
388
+ "epoch": 0.6,
389
+ "eval_loss": 2.166550874710083,
390
+ "eval_runtime": 2.9777,
391
+ "eval_samples_per_second": 3.358,
392
+ "eval_steps_per_second": 0.672,
393
+ "step": 60
394
+ },
395
+ {
396
+ "epoch": 0.61,
397
+ "learning_rate": 5.411329859370006e-05,
398
+ "loss": 2.0614,
399
+ "step": 61
400
+ },
401
+ {
402
+ "epoch": 0.62,
403
+ "learning_rate": 5.370903724953548e-05,
404
+ "loss": 2.061,
405
+ "step": 62
406
+ },
407
+ {
408
+ "epoch": 0.63,
409
+ "learning_rate": 5.329897756635779e-05,
410
+ "loss": 2.0868,
411
+ "step": 63
412
+ },
413
+ {
414
+ "epoch": 0.64,
415
+ "learning_rate": 5.2883231650389946e-05,
416
+ "loss": 2.0659,
417
+ "step": 64
418
+ },
419
+ {
420
+ "epoch": 0.65,
421
+ "learning_rate": 5.2461913162414206e-05,
422
+ "loss": 2.1325,
423
+ "step": 65
424
+ },
425
+ {
426
+ "epoch": 0.66,
427
+ "learning_rate": 5.203513728669831e-05,
428
+ "loss": 2.0955,
429
+ "step": 66
430
+ },
431
+ {
432
+ "epoch": 0.67,
433
+ "learning_rate": 5.160302069950537e-05,
434
+ "loss": 1.9692,
435
+ "step": 67
436
+ },
437
+ {
438
+ "epoch": 0.68,
439
+ "learning_rate": 5.116568153719566e-05,
440
+ "loss": 2.0858,
441
+ "step": 68
442
+ },
443
+ {
444
+ "epoch": 0.69,
445
+ "learning_rate": 5.072323936392936e-05,
446
+ "loss": 2.0884,
447
+ "step": 69
448
+ },
449
+ {
450
+ "epoch": 0.7,
451
+ "learning_rate": 5.027581513897888e-05,
452
+ "loss": 1.952,
453
+ "step": 70
454
+ },
455
+ {
456
+ "epoch": 0.71,
457
+ "learning_rate": 4.9823531183659796e-05,
458
+ "loss": 2.0223,
459
+ "step": 71
460
+ },
461
+ {
462
+ "epoch": 0.72,
463
+ "learning_rate": 4.936651114788949e-05,
464
+ "loss": 2.0098,
465
+ "step": 72
466
+ },
467
+ {
468
+ "epoch": 0.73,
469
+ "learning_rate": 4.890487997638242e-05,
470
+ "loss": 1.9965,
471
+ "step": 73
472
+ },
473
+ {
474
+ "epoch": 0.74,
475
+ "learning_rate": 4.843876387449151e-05,
476
+ "loss": 2.1097,
477
+ "step": 74
478
+ },
479
+ {
480
+ "epoch": 0.75,
481
+ "learning_rate": 4.796829027370489e-05,
482
+ "loss": 2.1051,
483
+ "step": 75
484
+ },
485
+ {
486
+ "epoch": 0.76,
487
+ "learning_rate": 4.749358779680735e-05,
488
+ "loss": 2.1672,
489
+ "step": 76
490
+ },
491
+ {
492
+ "epoch": 0.77,
493
+ "learning_rate": 4.701478622271621e-05,
494
+ "loss": 2.0641,
495
+ "step": 77
496
+ },
497
+ {
498
+ "epoch": 0.78,
499
+ "learning_rate": 4.653201645100101e-05,
500
+ "loss": 2.1899,
501
+ "step": 78
502
+ },
503
+ {
504
+ "epoch": 0.79,
505
+ "learning_rate": 4.6045410466096934e-05,
506
+ "loss": 2.007,
507
+ "step": 79
508
+ },
509
+ {
510
+ "epoch": 0.8,
511
+ "learning_rate": 4.555510130122151e-05,
512
+ "loss": 2.1355,
513
+ "step": 80
514
+ },
515
+ {
516
+ "epoch": 0.8,
517
+ "eval_loss": 2.160327434539795,
518
+ "eval_runtime": 2.9723,
519
+ "eval_samples_per_second": 3.364,
520
+ "eval_steps_per_second": 0.673,
521
+ "step": 80
522
+ },
523
+ {
524
+ "epoch": 0.81,
525
+ "learning_rate": 4.5061223002004694e-05,
526
+ "loss": 2.0046,
527
+ "step": 81
528
+ },
529
+ {
530
+ "epoch": 0.82,
531
+ "learning_rate": 4.456391058984217e-05,
532
+ "loss": 2.0902,
533
+ "step": 82
534
+ },
535
+ {
536
+ "epoch": 0.83,
537
+ "learning_rate": 4.406330002498178e-05,
538
+ "loss": 2.1259,
539
+ "step": 83
540
+ },
541
+ {
542
+ "epoch": 0.84,
543
+ "learning_rate": 4.3559528169353394e-05,
544
+ "loss": 1.9241,
545
+ "step": 84
546
+ },
547
+ {
548
+ "epoch": 0.85,
549
+ "learning_rate": 4.3052732749152215e-05,
550
+ "loss": 2.1537,
551
+ "step": 85
552
+ },
553
+ {
554
+ "epoch": 0.86,
555
+ "learning_rate": 4.254305231718579e-05,
556
+ "loss": 1.9482,
557
+ "step": 86
558
+ },
559
+ {
560
+ "epoch": 0.87,
561
+ "learning_rate": 4.2030626214995104e-05,
562
+ "loss": 2.0181,
563
+ "step": 87
564
+ },
565
+ {
566
+ "epoch": 0.88,
567
+ "learning_rate": 4.151559453475994e-05,
568
+ "loss": 1.9839,
569
+ "step": 88
570
+ },
571
+ {
572
+ "epoch": 0.89,
573
+ "learning_rate": 4.0998098080999136e-05,
574
+ "loss": 2.0735,
575
+ "step": 89
576
+ },
577
+ {
578
+ "epoch": 0.9,
579
+ "learning_rate": 4.047827833207597e-05,
580
+ "loss": 2.1691,
581
+ "step": 90
582
+ },
583
+ {
584
+ "epoch": 0.91,
585
+ "learning_rate": 3.995627740151942e-05,
586
+ "loss": 2.2761,
587
+ "step": 91
588
+ },
589
+ {
590
+ "epoch": 0.92,
591
+ "learning_rate": 3.9432237999171774e-05,
592
+ "loss": 2.1657,
593
+ "step": 92
594
+ },
595
+ {
596
+ "epoch": 0.93,
597
+ "learning_rate": 3.89063033921731e-05,
598
+ "loss": 1.9501,
599
+ "step": 93
600
+ },
601
+ {
602
+ "epoch": 0.94,
603
+ "learning_rate": 3.8378617365793576e-05,
604
+ "loss": 2.0869,
605
+ "step": 94
606
+ },
607
+ {
608
+ "epoch": 0.95,
609
+ "learning_rate": 3.784932418412385e-05,
610
+ "loss": 2.0112,
611
+ "step": 95
612
+ },
613
+ {
614
+ "epoch": 0.96,
615
+ "learning_rate": 3.731856855063486e-05,
616
+ "loss": 1.9212,
617
+ "step": 96
618
+ },
619
+ {
620
+ "epoch": 0.97,
621
+ "learning_rate": 3.678649556861724e-05,
622
+ "loss": 2.0395,
623
+ "step": 97
624
+ },
625
+ {
626
+ "epoch": 0.98,
627
+ "learning_rate": 3.625325070151155e-05,
628
+ "loss": 2.0179,
629
+ "step": 98
630
+ },
631
+ {
632
+ "epoch": 0.99,
633
+ "learning_rate": 3.571897973314004e-05,
634
+ "loss": 1.9713,
635
+ "step": 99
636
+ },
637
+ {
638
+ "epoch": 1.0,
639
+ "learning_rate": 3.51838287278508e-05,
640
+ "loss": 2.0756,
641
+ "step": 100
642
+ },
643
+ {
644
+ "epoch": 1.0,
645
+ "eval_loss": 2.1577982902526855,
646
+ "eval_runtime": 2.9721,
647
+ "eval_samples_per_second": 3.365,
648
+ "eval_steps_per_second": 0.673,
649
+ "step": 100
650
+ },
651
+ {
652
+ "epoch": 1.01,
653
+ "learning_rate": 3.4647943990585227e-05,
654
+ "loss": 1.9634,
655
+ "step": 101
656
+ },
657
+ {
658
+ "epoch": 1.02,
659
+ "learning_rate": 3.411147202687972e-05,
660
+ "loss": 2.1004,
661
+ "step": 102
662
+ },
663
+ {
664
+ "epoch": 1.03,
665
+ "learning_rate": 3.357455950281251e-05,
666
+ "loss": 1.9162,
667
+ "step": 103
668
+ },
669
+ {
670
+ "epoch": 1.04,
671
+ "learning_rate": 3.303735320490665e-05,
672
+ "loss": 1.9874,
673
+ "step": 104
674
+ },
675
+ {
676
+ "epoch": 1.05,
677
+ "learning_rate": 3.25e-05,
678
+ "loss": 2.0868,
679
+ "step": 105
680
+ },
681
+ {
682
+ "epoch": 1.06,
683
+ "learning_rate": 3.1962646795093345e-05,
684
+ "loss": 1.8818,
685
+ "step": 106
686
+ },
687
+ {
688
+ "epoch": 1.07,
689
+ "learning_rate": 3.142544049718748e-05,
690
+ "loss": 1.9699,
691
+ "step": 107
692
+ },
693
+ {
694
+ "epoch": 1.08,
695
+ "learning_rate": 3.088852797312027e-05,
696
+ "loss": 1.994,
697
+ "step": 108
698
+ },
699
+ {
700
+ "epoch": 1.09,
701
+ "learning_rate": 3.0352056009414768e-05,
702
+ "loss": 2.1169,
703
+ "step": 109
704
+ },
705
+ {
706
+ "epoch": 1.1,
707
+ "learning_rate": 2.98161712721492e-05,
708
+ "loss": 1.997,
709
+ "step": 110
710
+ },
711
+ {
712
+ "epoch": 1.11,
713
+ "learning_rate": 2.928102026685996e-05,
714
+ "loss": 2.1137,
715
+ "step": 111
716
+ },
717
+ {
718
+ "epoch": 1.12,
719
+ "learning_rate": 2.874674929848845e-05,
720
+ "loss": 2.0281,
721
+ "step": 112
722
+ },
723
+ {
724
+ "epoch": 1.13,
725
+ "learning_rate": 2.8213504431382755e-05,
726
+ "loss": 2.0649,
727
+ "step": 113
728
+ },
729
+ {
730
+ "epoch": 1.14,
731
+ "learning_rate": 2.768143144936513e-05,
732
+ "loss": 1.9891,
733
+ "step": 114
734
+ },
735
+ {
736
+ "epoch": 1.15,
737
+ "learning_rate": 2.715067581587614e-05,
738
+ "loss": 1.9841,
739
+ "step": 115
740
+ },
741
+ {
742
+ "epoch": 1.16,
743
+ "learning_rate": 2.6621382634206436e-05,
744
+ "loss": 2.1146,
745
+ "step": 116
746
+ },
747
+ {
748
+ "epoch": 1.17,
749
+ "learning_rate": 2.6093696607826892e-05,
750
+ "loss": 2.0479,
751
+ "step": 117
752
+ },
753
+ {
754
+ "epoch": 1.18,
755
+ "learning_rate": 2.556776200082823e-05,
756
+ "loss": 2.1255,
757
+ "step": 118
758
+ },
759
+ {
760
+ "epoch": 1.19,
761
+ "learning_rate": 2.5043722598480574e-05,
762
+ "loss": 2.0523,
763
+ "step": 119
764
+ },
765
+ {
766
+ "epoch": 1.2,
767
+ "learning_rate": 2.4521721667924026e-05,
768
+ "loss": 2.1502,
769
+ "step": 120
770
+ },
771
+ {
772
+ "epoch": 1.2,
773
+ "eval_loss": 2.1567764282226562,
774
+ "eval_runtime": 2.9942,
775
+ "eval_samples_per_second": 3.34,
776
+ "eval_steps_per_second": 0.668,
777
+ "step": 120
778
+ },
779
+ {
780
+ "epoch": 1.21,
781
+ "learning_rate": 2.4001901919000855e-05,
782
+ "loss": 2.0222,
783
+ "step": 121
784
+ },
785
+ {
786
+ "epoch": 1.22,
787
+ "learning_rate": 2.3484405465240055e-05,
788
+ "loss": 2.003,
789
+ "step": 122
790
+ },
791
+ {
792
+ "epoch": 1.23,
793
+ "learning_rate": 2.2969373785004897e-05,
794
+ "loss": 1.9935,
795
+ "step": 123
796
+ },
797
+ {
798
+ "epoch": 1.24,
799
+ "learning_rate": 2.2456947682814208e-05,
800
+ "loss": 1.92,
801
+ "step": 124
802
+ },
803
+ {
804
+ "epoch": 1.25,
805
+ "learning_rate": 2.1947267250847786e-05,
806
+ "loss": 2.0776,
807
+ "step": 125
808
+ },
809
+ {
810
+ "epoch": 1.26,
811
+ "learning_rate": 2.1440471830646597e-05,
812
+ "loss": 2.0785,
813
+ "step": 126
814
+ },
815
+ {
816
+ "epoch": 1.27,
817
+ "learning_rate": 2.093669997501821e-05,
818
+ "loss": 2.1597,
819
+ "step": 127
820
+ },
821
+ {
822
+ "epoch": 1.28,
823
+ "learning_rate": 2.0436089410157818e-05,
824
+ "loss": 2.0279,
825
+ "step": 128
826
+ },
827
+ {
828
+ "epoch": 1.29,
829
+ "learning_rate": 1.993877699799531e-05,
830
+ "loss": 2.107,
831
+ "step": 129
832
+ },
833
+ {
834
+ "epoch": 1.3,
835
+ "learning_rate": 1.9444898698778497e-05,
836
+ "loss": 2.121,
837
+ "step": 130
838
+ },
839
+ {
840
+ "epoch": 1.31,
841
+ "learning_rate": 1.8954589533903067e-05,
842
+ "loss": 2.1024,
843
+ "step": 131
844
+ },
845
+ {
846
+ "epoch": 1.32,
847
+ "learning_rate": 1.8467983548998984e-05,
848
+ "loss": 2.1763,
849
+ "step": 132
850
+ },
851
+ {
852
+ "epoch": 1.33,
853
+ "learning_rate": 1.798521377728379e-05,
854
+ "loss": 2.0075,
855
+ "step": 133
856
+ },
857
+ {
858
+ "epoch": 1.34,
859
+ "learning_rate": 1.750641220319265e-05,
860
+ "loss": 2.0791,
861
+ "step": 134
862
+ },
863
+ {
864
+ "epoch": 1.35,
865
+ "learning_rate": 1.703170972629511e-05,
866
+ "loss": 2.1985,
867
+ "step": 135
868
+ },
869
+ {
870
+ "epoch": 1.36,
871
+ "learning_rate": 1.6561236125508495e-05,
872
+ "loss": 2.1254,
873
+ "step": 136
874
+ },
875
+ {
876
+ "epoch": 1.37,
877
+ "learning_rate": 1.6095120023617585e-05,
878
+ "loss": 2.0442,
879
+ "step": 137
880
+ },
881
+ {
882
+ "epoch": 1.38,
883
+ "learning_rate": 1.5633488852110516e-05,
884
+ "loss": 2.0451,
885
+ "step": 138
886
+ },
887
+ {
888
+ "epoch": 1.39,
889
+ "learning_rate": 1.5176468816340209e-05,
890
+ "loss": 2.029,
891
+ "step": 139
892
+ },
893
+ {
894
+ "epoch": 1.4,
895
+ "learning_rate": 1.472418486102113e-05,
896
+ "loss": 2.108,
897
+ "step": 140
898
+ },
899
+ {
900
+ "epoch": 1.4,
901
+ "eval_loss": 2.153236150741577,
902
+ "eval_runtime": 2.9714,
903
+ "eval_samples_per_second": 3.365,
904
+ "eval_steps_per_second": 0.673,
905
+ "step": 140
906
+ },
907
+ {
908
+ "epoch": 1.41,
909
+ "learning_rate": 1.427676063607064e-05,
910
+ "loss": 2.0035,
911
+ "step": 141
912
+ },
913
+ {
914
+ "epoch": 1.42,
915
+ "learning_rate": 1.3834318462804333e-05,
916
+ "loss": 2.0776,
917
+ "step": 142
918
+ },
919
+ {
920
+ "epoch": 1.43,
921
+ "learning_rate": 1.3396979300494626e-05,
922
+ "loss": 1.9116,
923
+ "step": 143
924
+ },
925
+ {
926
+ "epoch": 1.44,
927
+ "learning_rate": 1.2964862713301684e-05,
928
+ "loss": 2.0144,
929
+ "step": 144
930
+ },
931
+ {
932
+ "epoch": 1.45,
933
+ "learning_rate": 1.2538086837585795e-05,
934
+ "loss": 2.1723,
935
+ "step": 145
936
+ },
937
+ {
938
+ "epoch": 1.46,
939
+ "learning_rate": 1.2116768349610045e-05,
940
+ "loss": 2.0467,
941
+ "step": 146
942
+ },
943
+ {
944
+ "epoch": 1.47,
945
+ "learning_rate": 1.1701022433642207e-05,
946
+ "loss": 1.9612,
947
+ "step": 147
948
+ },
949
+ {
950
+ "epoch": 1.48,
951
+ "learning_rate": 1.1290962750464517e-05,
952
+ "loss": 2.0506,
953
+ "step": 148
954
+ },
955
+ {
956
+ "epoch": 1.49,
957
+ "learning_rate": 1.0886701406299927e-05,
958
+ "loss": 1.9718,
959
+ "step": 149
960
+ },
961
+ {
962
+ "epoch": 1.5,
963
+ "learning_rate": 1.048834892216342e-05,
964
+ "loss": 2.0387,
965
+ "step": 150
966
+ }
967
+ ],
968
+ "logging_steps": 1,
969
+ "max_steps": 200,
970
+ "num_train_epochs": 2,
971
+ "save_steps": 50,
972
+ "total_flos": 1.251428724071465e+17,
973
+ "trial_name": null,
974
+ "trial_params": null
975
+ }
prompt-a/twochar/checkpoint-150/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9212c1ed1ebf30dd523c0f488bee8e1daf66e67eaaf76d60029260d57f02f88
3
+ size 4411
prompt-a/twochar/checkpoint-200/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-200/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-200/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3686a6854a70ed7424cb6434eaf2d2d208980a99034b0d2075128de53b3297a2
3
+ size 125374989
prompt-a/twochar/checkpoint-200/adapter_model/README.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+
7
+ The following `bitsandbytes` quantization config was used during training:
8
+ - quant_method: bitsandbytes
9
+ - load_in_8bit: False
10
+ - load_in_4bit: True
11
+ - llm_int8_threshold: 6.0
12
+ - llm_int8_skip_modules: None
13
+ - llm_int8_enable_fp32_cpu_offload: False
14
+ - llm_int8_has_fp16_weight: False
15
+ - bnb_4bit_quant_type: nf4
16
+ - bnb_4bit_use_double_quant: True
17
+ - bnb_4bit_compute_dtype: bfloat16
18
+ ### Framework versions
19
+
20
+
21
+ - PEFT 0.6.0.dev0
prompt-a/twochar/checkpoint-200/adapter_model/adapter_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "meta-llama/Llama-2-13b-hf",
4
+ "bias": "none",
5
+ "fan_in_fan_out": null,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.01,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 8,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "down_proj",
18
+ "k_proj",
19
+ "o_proj",
20
+ "v_proj",
21
+ "up_proj",
22
+ "q_proj",
23
+ "gate_proj"
24
+ ],
25
+ "task_type": "CAUSAL_LM"
26
+ }
prompt-a/twochar/checkpoint-200/adapter_model/adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3686a6854a70ed7424cb6434eaf2d2d208980a99034b0d2075128de53b3297a2
3
+ size 125374989
prompt-a/twochar/checkpoint-200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85cc40f0d7f81e4b3136a22c9ee61293c7e2a46795fded91d441e599b632fe7b
3
+ size 250681597
prompt-a/twochar/checkpoint-200/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb71352271decccd21025740fb7c4585441f353b8cbc20f09b524c2fe7cf027d
3
+ size 21687
prompt-a/twochar/checkpoint-200/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd2535ba0a236e9acf8608fce0f4038a69c182287799399a90be854fb25b7c34
3
+ size 21687
prompt-a/twochar/checkpoint-200/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eba2c62c4b80d00ace4ba0acc8ebdad176eddf3cd3a5b1050f7df7c4700163a6
3
+ size 21687