Upload folder using huggingface_hub

README.md

@@ -6,28 +6,12 @@ tags:
   - transformer
   - causal-lm
   - world-model
-datasets:
-  - random-self-play
-model-index:
-  - name: pawn-small
-    results:
-      - task:
-          type: next-move-prediction
-        metrics:
-          - name: Val Loss
-            type: loss
-            value: 3.15
-          - name: Val Accuracy
-            type: accuracy
-            value: 6.7
 ---
 
 # PAWN-SMALL
 
 A causal transformer trained on random chess games, designed as a testbed for finetuning and augmentation methods at small scales.
 
-## Model Details
-
 | | |
 |---|---|
 | **Parameters** | 9.5M |
@@ -35,31 +19,23 @@ A causal transformer trained on random chess games, designed as a testbed for fi
 | **d_model** | 256 |
 | **Layers** | 8 |
 | **Heads** | 4 |
-| **Vocabulary** | 4,278 tokens (4,096 grid + 176 promotions + 5 outcomes + 1 PAD) |
-| **Sequence length** | 256 |
-| **Training steps** | 80K/100K |
-| **Best val loss** | 3.150 (step 80,000) |
+| **Best val loss** | 3.1500 (step 81,000) |
 | **Best val accuracy** | 6.7% |
 
 ## Usage
 
 ```python
-import torch
+from safetensors.torch import load_file
 from pawn.config import CLMConfig
 from pawn.model import PAWNCLM
 
 cfg = CLMConfig.small()
 model = PAWNCLM(cfg)
-
-ckpt = torch.load("model.pt", map_location="cpu", weights_only=False)
-model.load_state_dict(ckpt["model_state_dict"])
+model.load_state_dict(load_file("model.safetensors"))
 model.eval()
 ```
 
-## Training
-
-Trained from scratch on random self-play games generated by a Rust chess engine (shakmaty).
-See the [PAWN repository](https://github.com/thomas-schweich/PAWN) for training code, data pipeline, and evaluation suite.
+See the [PAWN repository](https://github.com/thomas-schweich/PAWN) for training code and evaluation suite.
 
 ## License
 

config.json

@@ -0,0 +1,39 @@
+{
+  "format_version": 1,
+  "checkpoint_type": "pretrain",
+  "model_config": {
+    "vocab_size": 4278,
+    "max_seq_len": 256,
+    "n_outcomes": 5,
+    "d_model": 256,
+    "n_layers": 8,
+    "n_heads": 4,
+    "d_ff": 1024,
+    "dropout": 0.0,
+    "rope_base": 10000.0
+  },
+  "training_config": {
+    "lr": 0.0003,
+    "weight_decay": 0.01,
+    "max_grad_norm": 1.0,
+    "warmup_steps": 1000,
+    "total_steps": 100000,
+    "batch_size": 256,
+    "max_ply": 256,
+    "discard_ply_limit": false,
+    "num_workers": 4,
+    "use_amp": true,
+    "accumulation_steps": 1,
+    "log_interval": 10,
+    "eval_interval": 500,
+    "checkpoint_interval": 5000,
+    "base_seed": 42,
+    "val_seed": 9223372036854775807,
+    "val_games": 512,
+    "checkpoint_dir": "logs/run_20260322_194632/checkpoints",
+    "log_dir": "logs",
+    "use_wandb": false,
+    "wandb_project": "pawn",
+    "device": "cuda"
+  }
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e376e1fb26da29408a2b65f64b177417b2692b32d1f1ecab2b9973c224c9b29
+size 38102280

optimizer.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c48cda6c937c0572ea98abc2ecb9708bd9b47c972b7dc11c2ffc4b588fe1cba
+size 76210148

training_state.json

@@ -0,0 +1,118 @@
+{
+  "format_version": 1,
+  "global_step": 80000,
+  "scheduler_state_dict": {
+    "step": 80000
+  },
+  "scaler_state_dict": {
+    "scale": 2097152.0,
+    "growth_factor": 2.0,
+    "backoff_factor": 0.5,
+    "growth_interval": 2000,
+    "_growth_tracker": 19
+  },
+  "optimizer_meta": {
+    "param_groups": [
+      {
+        "lr": 5.628851523330708e-05,
+        "betas": [
+          0.9,
+          0.95
+        ],
+        "eps": 1e-08,
+        "weight_decay": 0.01,
+        "amsgrad": false,
+        "maximize": false,
+        "foreach": null,
+        "capturable": false,
+        "differentiable": false,
+        "fused": null,
+        "decoupled_weight_decay": true,
+        "params": [
+          0,
+          1,
+          2,
+          3,
+          4,
+          5,
+          6,
+          7,
+          8,
+          9,
+          10,
+          11,
+          12,
+          13,
+          14,
+          15,
+          16,
+          17,
+          18,
+          19,
+          20,
+          21,
+          22,
+          23,
+          24,
+          25,
+          26,
+          27,
+          28,
+          29,
+          30,
+          31,
+          32,
+          33,
+          34,
+          35,
+          36,
+          37,
+          38,
+          39,
+          40,
+          41,
+          42,
+          43,
+          44,
+          45,
+          46,
+          47,
+          48,
+          49,
+          50,
+          51,
+          52,
+          53,
+          54,
+          55,
+          56,
+          57,
+          58,
+          59,
+          60,
+          61,
+          62,
+          63,
+          64,
+          65,
+          66,
+          67,
+          68,
+          69,
+          70,
+          71,
+          72,
+          73,
+          74,
+          75,
+          76,
+          77,
+          78
+        ]
+      }
+    ],
+    "scalars": null
+  },
+  "torch_rng_state": "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",
+  "cuda_rng_state": "nfeOSbJFGwAAAAAAAAAAAA=="
+}