End of training

README.md

@@ -1,9 +1,11 @@
 ---
 base_model: sohyunan/gemma-2-2b-it_controller_sft_random
+datasets: maze_5x5
 library_name: transformers
 model_name: gemma-2-2b-it_controller_sft_random_grpo
 tags:
 - generated_from_trainer
+- controller-grpo
 - trl
 - grpo
 licence: license
@@ -11,7 +13,7 @@ licence: license
 
 # Model Card for gemma-2-2b-it_controller_sft_random_grpo
 
-This model is a fine-tuned version of [sohyunan/gemma-2-2b-it_controller_sft_random](https://huggingface.co/sohyunan/gemma-2-2b-it_controller_sft_random).
+This model is a fine-tuned version of [sohyunan/gemma-2-2b-it_controller_sft_random](https://huggingface.co/sohyunan/gemma-2-2b-it_controller_sft_random) on the [maze_5x5](https://huggingface.co/datasets/maze_5x5) dataset.
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start

all_results.json

@@ -1,4 +1,9 @@
 {
+    "eval_loss": -2.9802322387695312e-08,
+    "eval_runtime": 200.6099,
+    "eval_samples": 2,
+    "eval_samples_per_second": 0.01,
+    "eval_steps_per_second": 0.005,
     "total_flos": 0.0,
     "train_loss": -4.1987126072247825e-08,
     "train_runtime": 410.9305,

config.json

@@ -31,6 +31,6 @@
   "sliding_window": 4096,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.49.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 256001
 }

eval_results.json

@@ -0,0 +1,7 @@
+{
+    "eval_loss": -2.9802322387695312e-08,
+    "eval_runtime": 200.6099,
+    "eval_samples": 2,
+    "eval_samples_per_second": 0.01,
+    "eval_steps_per_second": 0.005
+}