Hugging Face's logo

datafreak
/
laya-chess

Laya
Safetensors
chess
decision-model
Model card Files
xet
 Community
datafreak commited on
Commit
aca8ffb
·
verified ·
1 Parent(s): b0c3489

final: step 16145, 1033280 examples, eval {'val_loss': 1.3803, 'val_wp_mae': 0.091, 'top_move_acc': 0.23666666666666666, 'step': 16145}

Browse files
Files changed (4) hide show
  1. README.md +2 -2
  2. chess_meta.json +60 -56
  3. model.safetensors +1 -1
  4. train_state.pt +3 -0
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: `step15590` (step 15590, 997,760 training examples)
13
- - latest eval: `{'val_loss': 1.3893, 'val_wp_mae': 0.0943, 'top_move_acc': 0.23666666666666666, 'step': 14000}`
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 16145, 1,033,280 training examples)
13
+ - latest eval: `{'val_loss': 1.3803, 'val_wp_mae': 0.091, 'top_move_acc': 0.23666666666666666, 'step': 16145}`
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": "step15590",
3
- "final": false,
4
- "step": 15590,
5
- "consumed": 997760,
6
  "n_levels": 10,
7
  "crit": [
8
  "0-10%",
@@ -45,58 +45,6 @@
45
  "SEED": 0
46
  },
47
  "history": [
48
- {
49
- "step": 13150,
50
- "loss": 1.3983
51
- },
52
- {
53
- "step": 13200,
54
- "loss": 1.3956
55
- },
56
- {
57
- "step": 13250,
58
- "loss": 1.4094
59
- },
60
- {
61
- "step": 13300,
62
- "loss": 1.3916
63
- },
64
- {
65
- "step": 13350,
66
- "loss": 1.389
67
- },
68
- {
69
- "step": 13400,
70
- "loss": 1.3858
71
- },
72
- {
73
- "step": 13450,
74
- "loss": 1.4014
75
- },
76
- {
77
- "step": 13500,
78
- "loss": 1.3969
79
- },
80
- {
81
- "step": 13550,
82
- "loss": 1.3648
83
- },
84
- {
85
- "step": 13600,
86
- "loss": 1.4037
87
- },
88
- {
89
- "step": 13650,
90
- "loss": 1.4081
91
- },
92
- {
93
- "step": 13700,
94
- "loss": 1.4193
95
- },
96
- {
97
- "step": 13750,
98
- "loss": 1.4144
99
- },
100
  {
101
  "step": 13800,
102
  "loss": 1.4002
@@ -246,6 +194,62 @@
246
  {
247
  "step": 15550,
248
  "loss": 1.3914
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
249
  }
250
  ]
251
  }
 
1
  {
2
+ "tag": "final",
3
+ "final": true,
4
+ "step": 16145,
5
+ "consumed": 1033280,
6
  "n_levels": 10,
7
  "crit": [
8
  "0-10%",
 
45
  "SEED": 0
46
  },
47
  "history": [
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
48
  {
49
  "step": 13800,
50
  "loss": 1.4002
 
194
  {
195
  "step": 15550,
196
  "loss": 1.3914
197
+ },
198
+ {
199
+ "step": 15600,
200
+ "loss": 1.3927
201
+ },
202
+ {
203
+ "step": 15650,
204
+ "loss": 1.3722
205
+ },
206
+ {
207
+ "step": 15700,
208
+ "loss": 1.3652
209
+ },
210
+ {
211
+ "step": 15750,
212
+ "loss": 1.4063
213
+ },
214
+ {
215
+ "step": 15800,
216
+ "loss": 1.3709
217
+ },
218
+ {
219
+ "step": 15850,
220
+ "loss": 1.3887
221
+ },
222
+ {
223
+ "step": 15900,
224
+ "loss": 1.4116
225
+ },
226
+ {
227
+ "step": 15950,
228
+ "loss": 1.3723
229
+ },
230
+ {
231
+ "step": 16000,
232
+ "loss": 1.4105
233
+ },
234
+ {
235
+ "val_loss": 1.3828,
236
+ "val_wp_mae": 0.0915,
237
+ "top_move_acc": 0.23333333333333334,
238
+ "step": 16000
239
+ },
240
+ {
241
+ "step": 16050,
242
+ "loss": 1.3975
243
+ },
244
+ {
245
+ "step": 16100,
246
+ "loss": 1.3831
247
+ },
248
+ {
249
+ "val_loss": 1.3803,
250
+ "val_wp_mae": 0.091,
251
+ "top_move_acc": 0.23666666666666666,
252
+ "step": 16145
253
  }
254
  ]
255
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6d23e0bb07b1ba199d13aaa2ca10048a7bbf348083ba5415e33372dd3e51a6a8
3
  size 842609220
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ec7a035dc1cd8457aa3f39cfe6d8b30478146d27e9ce5091532bb3bd449a81b
3
  size 842609220
train_state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4eb740d8496f0832f9e622f430d8fe21b960bf9ac60604020596f13c63d8af96
3
+ size 3370524709