Chess Challenge submission by Bichrai

README.md

@@ -14,13 +14,13 @@ Chess model submitted to the LLM Course Chess Challenge.
 ## Submission Info
 
 - **Submitted by**: [Bichrai](https://huggingface.co/Bichrai)
-- **Parameters**: 821,296
+- **Parameters**: 876,544
 - **Organization**: LLM-course
 
 ## Model Details
 
 - **Architecture**: Chess Transformer (GPT-style)
 - **Vocab size**: 1682
-- **Embedding dim**: 112
+- **Embedding dim**: 128
 - **Layers**: 5
 - **Heads**: 4

config.json

@@ -6,11 +6,12 @@
   "dropout": 0.1,
   "dtype": "float32",
   "eos_token_id": 2,
+  "layer_norm_epsilon": 1e-05,
   "model_type": "chess_transformer",
   "n_ctx": 256,
-  "n_embd": 112,
+  "n_embd": 128,
   "n_head": 4,
-  "n_inner": 336,
+  "n_inner": 256,
   "n_layer": 5,
   "pad_token_id": 0,
   "rms_norm_eps": 1e-06,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5f2ecf2379e1f892827eb5d85dc3f5422614d4e0b1d66ce350e6928b00bce750
-size 3289680
+oid sha256:13c9467fbfc32654d0a4b656df524a9a0010009519786db73d3fa8e48847aba2
+size 3510672