End of training

README.md

@@ -2,9 +2,16 @@
 license: apache-2.0
 base_model: alignment-handbook/mistral-7b-sft-constitutional-ai
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
+- HuggingFaceH4/cai-conversation-harmless
 model-index:
 - name: mistral-7b-dpo-constitutional-ai
   results: []
@@ -15,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral-7b-dpo-constitutional-ai
 
-This model is a fine-tuned version of [alignment-handbook/mistral-7b-sft-constitutional-ai](https://huggingface.co/alignment-handbook/mistral-7b-sft-constitutional-ai) on the None dataset.
+This model is a fine-tuned version of [alignment-handbook/mistral-7b-sft-constitutional-ai](https://huggingface.co/alignment-handbook/mistral-7b-sft-constitutional-ai) on the HuggingFaceH4/ultrafeedback_binarized and the HuggingFaceH4/cai-conversation-harmless datasets.
 It achieves the following results on the evaluation set:
-- Loss: 0.5467
-- Rewards/chosen: -3.4828
-- Rewards/rejected: -6.0365
-- Rewards/accuracies: 0.7025
-- Rewards/margins: 2.5536
-- Logps/rejected: -230.1307
-- Logps/chosen: -225.5250
-- Logits/rejected: -2.9599
-- Logits/chosen: -2.9763
+- Loss: 0.6064
+- Rewards/chosen: -6.2332
+- Rewards/rejected: -9.9320
+- Rewards/accuracies: 0.6825
+- Rewards/margins: 3.6988
+- Logps/rejected: -269.0856
+- Logps/chosen: -253.0284
+- Logits/rejected: -2.8561
+- Logits/chosen: -2.8752
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 3.0,
+    "eval_logits/chosen": -2.8751940727233887,
+    "eval_logits/rejected": -2.8561081886291504,
+    "eval_logps/chosen": -253.02841186523438,
+    "eval_logps/rejected": -269.08563232421875,
+    "eval_loss": 0.6063873767852783,
+    "eval_rewards/accuracies": 0.6825000047683716,
+    "eval_rewards/chosen": -6.233180522918701,
+    "eval_rewards/margins": 3.6988027095794678,
+    "eval_rewards/rejected": -9.931983947753906,
+    "eval_runtime": 168.7249,
+    "eval_samples": 3156,
+    "eval_samples_per_second": 18.705,
+    "eval_steps_per_second": 0.296,
     "train_loss": 0.25373340110176973,
     "train_runtime": 3870.3099,
     "train_samples": 8239,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 3.0,
+    "eval_logits/chosen": -2.8751940727233887,
+    "eval_logits/rejected": -2.8561081886291504,
+    "eval_logps/chosen": -253.02841186523438,
+    "eval_logps/rejected": -269.08563232421875,
+    "eval_loss": 0.6063873767852783,
+    "eval_rewards/accuracies": 0.6825000047683716,
+    "eval_rewards/chosen": -6.233180522918701,
+    "eval_rewards/margins": 3.6988027095794678,
+    "eval_rewards/rejected": -9.931983947753906,
+    "eval_runtime": 168.7249,
+    "eval_samples": 3156,
+    "eval_samples_per_second": 18.705,
+    "eval_steps_per_second": 0.296
+}

runs/Apr10_00-26-56_gx-dev-master-0/events.out.tfevents.1712713251.gx-dev-master-0.951.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9944e36f17a578cf2a48c1be62f92f1a512c01e7ddc71d679663d85fb1a408c
+size 780