End of training

README.md

@@ -2,10 +2,13 @@
 license: gemma
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - sft
 - generated_from_trainer
 base_model: google/gemma-7b
+datasets:
+- preference-data
 model-index:
 - name: downstream_gemma
   results: []
@@ -16,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # downstream_gemma
 
-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 preference-data dataset.
 
 ## Model description
 

all_results.json

@@ -1,5 +1,9 @@
 {
     "epoch": 1.0,
+    "eval_runtime": 3.7697,
+    "eval_samples": 10,
+    "eval_samples_per_second": 2.653,
+    "eval_steps_per_second": 0.796,
     "total_flos": 1.3011067827388416e+16,
     "train_loss": 1.2484874595598594,
     "train_runtime": 22644.1998,

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "google/gemma-7b",
+  "architectures": [
+    "GemmaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 2,
+  "eos_token_id": 1,
+  "head_dim": 256,
+  "hidden_act": "gelu",
+  "hidden_activation": null,
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 24576,
+  "max_position_embeddings": 8192,
+  "model_type": "gemma",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 16,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.44.2",
+  "use_cache": true,
+  "vocab_size": 256000
+}

eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 1.0,
+    "eval_runtime": 3.7697,
+    "eval_samples": 10,
+    "eval_samples_per_second": 2.653,
+    "eval_steps_per_second": 0.796
+}