Hugging Face's logo

datafreak
/
laya-chess

Laya
Safetensors
chess
decision-model
Model card Files
xet
 Community
datafreak commited on
Commit
639e575
·
verified ·
1 Parent(s): 478dc3a

final: step 32000, 2048000 examples, eval {'val_loss': 1.3315, 'val_wp_mae': 0.082, 'top_move_acc': 0.26666666666666666, 'step': 32000}

Browse files
Files changed (4) hide show
  1. README.md +2 -2
  2. chess_meta.json +68 -64
  3. model.safetensors +1 -1
  4. train_state.pt +1 -1
README.md CHANGED
@@ -9,7 +9,7 @@ tags: [chess, laya, decision-model]
9
  for each legal move, a Laya `score` question predicts the side to move's win chance over 10 levels.
10
  The engine plays the move with the highest expected win chance (no search).
11
 
12
- - checkpoint: `step31388` (step 31388, 2,008,832 training examples)
13
- - latest eval: `{'val_loss': 1.3347, 'val_wp_mae': 0.0836, 'top_move_acc': 0.26666666666666666, 'step': 30000}`
14
  - encoding / levels / prompt template: see `chess_meta.json`
15
  - code: github.com/devroopsaha744/LayaChess
 
9
  for each legal move, a Laya `score` question predicts the side to move's win chance over 10 levels.
10
  The engine plays the move with the highest expected win chance (no search).
11
 
12
+ - checkpoint: `final` (step 32000, 2,048,000 training examples)
13
+ - latest eval: `{'val_loss': 1.3315, 'val_wp_mae': 0.082, 'top_move_acc': 0.26666666666666666, 'step': 32000}`
14
  - encoding / levels / prompt template: see `chess_meta.json`
15
  - code: github.com/devroopsaha744/LayaChess
chess_meta.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "tag": "step31388",
3
- "final": false,
4
- "step": 31388,
5
- "consumed": 2008832,
6
  "n_levels": 10,
7
  "crit": [
8
  "0-10%",
@@ -45,66 +45,6 @@
45
  "SEED": 0
46
  },
47
  "history": [
48
- {
49
- "step": 28950,
50
- "loss": 1.3458
51
- },
52
- {
53
- "step": 29000,
54
- "loss": 1.3656
55
- },
56
- {
57
- "step": 29050,
58
- "loss": 1.3391
59
- },
60
- {
61
- "step": 29100,
62
- "loss": 1.356
63
- },
64
- {
65
- "step": 29150,
66
- "loss": 1.3411
67
- },
68
- {
69
- "step": 29200,
70
- "loss": 1.3312
71
- },
72
- {
73
- "step": 29250,
74
- "loss": 1.3207
75
- },
76
- {
77
- "step": 29300,
78
- "loss": 1.3497
79
- },
80
- {
81
- "step": 29350,
82
- "loss": 1.3586
83
- },
84
- {
85
- "step": 29400,
86
- "loss": 1.3554
87
- },
88
- {
89
- "step": 29450,
90
- "loss": 1.3467
91
- },
92
- {
93
- "step": 29500,
94
- "loss": 1.3324
95
- },
96
- {
97
- "step": 29550,
98
- "loss": 1.3634
99
- },
100
- {
101
- "step": 29600,
102
- "loss": 1.3384
103
- },
104
- {
105
- "step": 29650,
106
- "loss": 1.3214
107
- },
108
  {
109
  "step": 29700,
110
  "loss": 1.3451
@@ -246,6 +186,70 @@
246
  {
247
  "step": 31350,
248
  "loss": 1.3401
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
249
  }
250
  ]
251
  }
 
1
  {
2
+ "tag": "final",
3
+ "final": true,
4
+ "step": 32000,
5
+ "consumed": 2048000,
6
  "n_levels": 10,
7
  "crit": [
8
  "0-10%",
 
45
  "SEED": 0
46
  },
47
  "history": [
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
48
  {
49
  "step": 29700,
50
  "loss": 1.3451
 
186
  {
187
  "step": 31350,
188
  "loss": 1.3401
189
+ },
190
+ {
191
+ "step": 31400,
192
+ "loss": 1.36
193
+ },
194
+ {
195
+ "step": 31450,
196
+ "loss": 1.3304
197
+ },
198
+ {
199
+ "step": 31500,
200
+ "loss": 1.3174
201
+ },
202
+ {
203
+ "step": 31550,
204
+ "loss": 1.3442
205
+ },
206
+ {
207
+ "step": 31600,
208
+ "loss": 1.3621
209
+ },
210
+ {
211
+ "step": 31650,
212
+ "loss": 1.3217
213
+ },
214
+ {
215
+ "step": 31700,
216
+ "loss": 1.3186
217
+ },
218
+ {
219
+ "step": 31750,
220
+ "loss": 1.329
221
+ },
222
+ {
223
+ "step": 31800,
224
+ "loss": 1.333
225
+ },
226
+ {
227
+ "step": 31850,
228
+ "loss": 1.3585
229
+ },
230
+ {
231
+ "step": 31900,
232
+ "loss": 1.3238
233
+ },
234
+ {
235
+ "step": 31950,
236
+ "loss": 1.3538
237
+ },
238
+ {
239
+ "step": 32000,
240
+ "loss": 1.3216
241
+ },
242
+ {
243
+ "val_loss": 1.3315,
244
+ "val_wp_mae": 0.082,
245
+ "top_move_acc": 0.26666666666666666,
246
+ "step": 32000
247
+ },
248
+ {
249
+ "val_loss": 1.3315,
250
+ "val_wp_mae": 0.082,
251
+ "top_move_acc": 0.26666666666666666,
252
+ "step": 32000
253
  }
254
  ]
255
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:89d47469af205cc17ece4ca8af6703536aa983f6f74437c442dfcb0f3621a378
3
  size 842609220
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3eeb55042a463bbbd3b159bb0e7d9833135d57248bf7faf78ef5b18b69c1bd68
3
  size 842609220
train_state.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4eb740d8496f0832f9e622f430d8fe21b960bf9ac60604020596f13c63d8af96
3
  size 3370524709
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:310adc86bd9605418b82b2f2ec8bbaa3bd9b256bb987ecb1292bc8c41cc5f7c5
3
  size 3370524709