End of training

README.md

@@ -2,12 +2,13 @@
 license: apache-2.0
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
 base_model: mistralai/Mistral-7B-v0.1
 datasets:
-- generator
+- HuggingFaceH4/ultrachat_200k
 model-index:
 - name: zephyr-7b-sft-qlora
   results: []
@@ -18,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-sft-qlora
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the HuggingFaceH4/ultrachat_200k dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.9517
 

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 1.0,
+    "eval_loss": 0.9517214298248291,
+    "eval_runtime": 970.177,
+    "eval_samples": 23109,
+    "eval_samples_per_second": 15.905,
+    "eval_steps_per_second": 0.995,
     "train_loss": 0.8657062510599203,
     "train_runtime": 31020.7888,
     "train_samples": 207864,

config.json

@@ -18,7 +18,8 @@
   "quantization_config": {
     "_load_in_4bit": true,
     "_load_in_8bit": false,
-    "bnb_4bit_compute_dtype": "float16",
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
     "bnb_4bit_quant_type": "nf4",
     "bnb_4bit_use_double_quant": false,
     "llm_int8_enable_fp32_cpu_offload": false,
@@ -33,8 +34,8 @@
   "rope_theta": 10000.0,
   "sliding_window": 4096,
   "tie_word_embeddings": false,
-  "torch_dtype": "float16",
-  "transformers_version": "4.38.2",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.3",
   "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.9498828649520874,
-    "eval_runtime": 985.6149,
-    "eval_samples": 23110,
-    "eval_samples_per_second": 15.656,
-    "eval_steps_per_second": 0.49
+    "eval_loss": 0.9517214298248291,
+    "eval_runtime": 970.177,
+    "eval_samples": 23109,
+    "eval_samples_per_second": 15.905,
+    "eval_steps_per_second": 0.995
 }