End of training

README.md

@@ -3,6 +3,7 @@ base_model: HuggingFaceH4/mistral-7b-sft-beta
 library_name: peft
 license: mit
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
@@ -18,15 +19,15 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.5084
-- Rewards/chosen: -0.0008
-- Rewards/rejected: -0.9080
+- Loss: 0.5082
+- Rewards/chosen: 0.0025
+- Rewards/rejected: -0.9047
 - Rewards/accuracies: 0.7222
 - Rewards/margins: 0.9072
-- Logps/rejected: -276.7181
-- Logps/chosen: -271.8789
-- Logits/rejected: -2.7174
-- Logits/chosen: -2.7372
+- Logps/rejected: -276.6852
+- Logps/chosen: -271.8461
+- Logits/rejected: -2.7167
+- Logits/chosen: -2.7365
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9997382884061764,
+    "eval_logits/chosen": -2.736534833908081,
+    "eval_logits/rejected": -2.716714382171631,
+    "eval_logps/chosen": -271.84613037109375,
+    "eval_logps/rejected": -276.6851501464844,
+    "eval_loss": 0.5081929564476013,
+    "eval_rewards/accuracies": 0.7222222089767456,
+    "eval_rewards/chosen": 0.002502653282135725,
+    "eval_rewards/margins": 0.9071723222732544,
+    "eval_rewards/rejected": -0.9046697020530701,
+    "eval_runtime": 352.642,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 5.671,
+    "eval_steps_per_second": 0.179,
     "total_flos": 0.0,
     "train_loss": 0.5278735879828168,
     "train_runtime": 28474.7268,

config.json

@@ -1,11 +1,12 @@
 {
-  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "_name_or_path": "HuggingFaceH4/mistral-7b-sft-beta",
   "architectures": [
     "MistralForCausalLM"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 1,
   "eos_token_id": 2,
+  "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 4096,
   "initializer_range": 0.02,
@@ -20,7 +21,7 @@
   "sliding_window": 4096,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.38.2",
+  "transformers_version": "4.45.2",
   "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 1.0,
-    "eval_logits/chosen": 0.5316108465194702,
-    "eval_logits/rejected": 0.9396071434020996,
-    "eval_logps/chosen": -447.05303955078125,
-    "eval_logps/rejected": -536.2667236328125,
-    "eval_loss": 0.49303239583969116,
-    "eval_rewards/accuracies": 0.7242063283920288,
-    "eval_rewards/chosen": -1.7955820560455322,
-    "eval_rewards/margins": 0.9434418678283691,
-    "eval_rewards/rejected": -2.7390236854553223,
-    "eval_runtime": 363.1163,
+    "epoch": 0.9997382884061764,
+    "eval_logits/chosen": -2.736534833908081,
+    "eval_logits/rejected": -2.716714382171631,
+    "eval_logps/chosen": -271.84613037109375,
+    "eval_logps/rejected": -276.6851501464844,
+    "eval_loss": 0.5081929564476013,
+    "eval_rewards/accuracies": 0.7222222089767456,
+    "eval_rewards/chosen": 0.002502653282135725,
+    "eval_rewards/margins": 0.9071723222732544,
+    "eval_rewards/rejected": -0.9046697020530701,
+    "eval_runtime": 352.642,
     "eval_samples": 2000,
-    "eval_samples_per_second": 5.508,
-    "eval_steps_per_second": 0.173
+    "eval_samples_per_second": 5.671,
+    "eval_steps_per_second": 0.179
 }