Training in progress, step 100

README.md

@@ -3,9 +3,12 @@ library_name: transformers
 license: gemma
 base_model: google/gemma-7b
 tags:
+- alignment-handbook
 - trl
 - orpo
 - generated_from_trainer
+datasets:
+- argilla/dpo-mix-7k
 model-index:
 - name: gemma-7b-orpo
   results: []
@@ -16,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma-7b-orpo
 
-This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on an unknown dataset.
+This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the argilla/dpo-mix-7k dataset.
 It achieves the following results on the evaluation set:
 - Loss: 1.4556
 - Rewards/chosen: -0.0513

all_results.json

@@ -1,5 +1,21 @@
 {
     "epoch": 1.0,
+    "eval_log_odds_chosen": 0.23976314067840576,
+    "eval_log_odds_ratio": -0.6928443908691406,
+    "eval_logits/chosen": 340.5321350097656,
+    "eval_logits/rejected": 312.9670104980469,
+    "eval_logps/chosen": -1.0267834663391113,
+    "eval_logps/rejected": -1.1786600351333618,
+    "eval_loss": 1.455647587776184,
+    "eval_nll_loss": 1.4095592498779297,
+    "eval_rewards/accuracies": 0.5107913613319397,
+    "eval_rewards/chosen": -0.051339175552129745,
+    "eval_rewards/margins": 0.0075938161462545395,
+    "eval_rewards/rejected": -0.058932989835739136,
+    "eval_runtime": 90.3296,
+    "eval_samples": 553,
+    "eval_samples_per_second": 6.122,
+    "eval_steps_per_second": 1.539,
     "total_flos": 0.0,
     "train_loss": 1.8019611810861456,
     "train_runtime": 4470.8327,

config.json

@@ -24,6 +24,6 @@
   "rope_theta": 10000.0,
   "torch_dtype": "float32",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 1.0,
+    "eval_log_odds_chosen": 0.23976314067840576,
+    "eval_log_odds_ratio": -0.6928443908691406,
+    "eval_logits/chosen": 340.5321350097656,
+    "eval_logits/rejected": 312.9670104980469,
+    "eval_logps/chosen": -1.0267834663391113,
+    "eval_logps/rejected": -1.1786600351333618,
+    "eval_loss": 1.455647587776184,
+    "eval_nll_loss": 1.4095592498779297,
+    "eval_rewards/accuracies": 0.5107913613319397,
+    "eval_rewards/chosen": -0.051339175552129745,
+    "eval_rewards/margins": 0.0075938161462545395,
+    "eval_rewards/rejected": -0.058932989835739136,
+    "eval_runtime": 90.3296,
+    "eval_samples": 553,
+    "eval_samples_per_second": 6.122,
+    "eval_steps_per_second": 1.539
+}

runs/Sep14_18-22-28_65ecb96dba42/events.out.tfevents.1726343212.65ecb96dba42.565148.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18a5b055d8d665befa123a3845689a9e1156773e068d00258536ae7316c2537f
+size 997

runs/Sep14_20-25-33_65ecb96dba42/events.out.tfevents.1726345592.65ecb96dba42.608261.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fc0990db8924114d337b7faf710da64f270f97325b7a282fe4b59fded640790
+size 122804

runs/Sep14_21-14-45_65ecb96dba42/events.out.tfevents.1726348544.65ecb96dba42.1985.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b288208fdc46fb8b73b249bc68e959a2a2c77d897ec30042299921bc09d410c
+size 23104

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:baf033bae0930cfe8415310102d9c022be625d63b2ff7efcb149c28652442604
+oid sha256:f046069520bac033c141ed4582bc738952f18218eb93a0db6c5507109a8824a5
 size 5560