End of training

README.md

@@ -2,12 +2,16 @@
 license: apache-2.0
 base_model: mistralai/Mistral-7B-Instruct-v0.2
 tags:
+- alignment-handbook
+- trl
+- sft
+- generated_from_trainer
 - trl
 - sft
 - alignment-handbook
 - generated_from_trainer
 datasets:
-- generator
+- preference-data
 model-index:
 - name: prometheus-7b-direct-0.1p-rare-seed42
   results: []
@@ -18,7 +22,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # prometheus-7b-direct-0.1p-rare-seed42
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the generator dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the preference-data dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.3326
 

all_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 0.9995119570522206,
     "eval_loss": 0.3325524628162384,
-    "eval_runtime": 97.3399,
+    "eval_runtime": 95.5011,
     "eval_samples": 1000,
-    "eval_samples_per_second": 3.103,
-    "eval_steps_per_second": 0.781,
+    "eval_samples_per_second": 3.162,
+    "eval_steps_per_second": 0.796,
     "total_flos": 214352422502400.0,
     "train_loss": 0.0,
     "train_runtime": 0.0086,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.43.4",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 0.9995119570522206,
     "eval_loss": 0.3325524628162384,
-    "eval_runtime": 97.3399,
+    "eval_runtime": 95.5011,
     "eval_samples": 1000,
-    "eval_samples_per_second": 3.103,
-    "eval_steps_per_second": 0.781
+    "eval_samples_per_second": 3.162,
+    "eval_steps_per_second": 0.796
 }

runs/Aug07_15-38-10_COE-CS-sv003/events.out.tfevents.1723045298.COE-CS-sv003.3042844.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b095ee3556a3aacc0b0746f3608343a80e0fb6d8956a10204394d0f606a00427
+size 359