End of training

README.md

@@ -2,11 +2,15 @@
 license: gemma
 base_model: google/gemma-2-2b-it
 tags:
+- alignment-handbook
+- trl
+- sft
+- generated_from_trainer
 - trl
 - sft
 - generated_from_trainer
 datasets:
-- generator
+- haoranxu/ALMA-Human-Parallel
 model-index:
 - name: gemma-2-2b-it-mt-sft-full_new
   results: []
@@ -17,7 +21,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # gemma-2-2b-it-mt-sft-full_new
 
-This model is a fine-tuned version of [google/gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it) on the generator dataset.
+This model is a fine-tuned version of [google/gemma-2-2b-it](https://huggingface.co/google/gemma-2-2b-it) on the haoranxu/ALMA-Human-Parallel dataset.
 It achieves the following results on the evaluation set:
 - Loss: 1.8862
 

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 0.9982363315696648,
+    "eval_loss": 1.8862216472625732,
+    "eval_runtime": 61.2833,
+    "eval_samples": 8016,
+    "eval_samples_per_second": 5.434,
+    "eval_steps_per_second": 2.725,
     "total_flos": 1.1264430584732058e+17,
     "train_loss": 1.8635691063142918,
     "train_runtime": 3704.3906,

config.json

@@ -31,6 +31,6 @@
   "sliding_window": 4096,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.43.3",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256000
 }

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.9982363315696648,
+    "eval_loss": 1.8862216472625732,
+    "eval_runtime": 61.2833,
+    "eval_samples": 8016,
+    "eval_samples_per_second": 5.434,
+    "eval_steps_per_second": 2.725
+}

runs/Aug02_00-26-24_poseidon/events.out.tfevents.1722562329.poseidon.1257643.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:841cd52a5f0f8ee8ad6577348bf7cf5f2d49ddd049739ac74fc8448f1ec1eba4
+size 359