Upload 5 _export models (75% memorize variants)
Browse files- cho-chikun_export/best_checkpoint.ckpt +3 -0
- cho-chikun_export/log.txt +420 -0
- cho-chikun_export/metadata.json +1 -0
- cho-chikun_export/model.bin.gz +3 -0
- kitani_export/best_checkpoint.ckpt +3 -0
- kitani_export/log.txt +420 -0
- kitani_export/metadata.json +1 -0
- kitani_export/model.bin.gz +3 -0
- nieweiping_export/best_checkpoint.ckpt +3 -0
- nieweiping_export/log.txt +482 -0
- nieweiping_export/metadata.json +1 -0
- nieweiping_export/model.bin.gz +3 -0
- shusaku_export/best_checkpoint.ckpt +3 -0
- shusaku_export/log.txt +420 -0
- shusaku_export/metadata.json +1 -0
- shusaku_export/model.bin.gz +3 -0
- takemiya_export/best_checkpoint.ckpt +3 -0
- takemiya_export/log.txt +420 -0
- takemiya_export/metadata.json +1 -0
- takemiya_export/model.bin.gz +3 -0
cho-chikun_export/best_checkpoint.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79cbd77a1b46bc34a66184d05ecff2a1948e5b0f89245791b68d43c0217f067d
|
| 3 |
+
size 214677715
|
cho-chikun_export/log.txt
ADDED
|
@@ -0,0 +1,420 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
['/home/ran/go-arm-rl/KataGo.OLD/python/export_model_pytorch.py', '-checkpoint', '/home/ran/go-arm-rl/models/cho-chikun_75/best_checkpoint.ckpt', '-export-dir', '/home/ran/go-arm-rl/models/cho-chikun_75', '-model-name', 'cho-chikun_75', '-filename-prefix', 'model', '-export-14-as-15']
|
| 2 |
+
{'version': 14, 'norm_kind': 'fixscaleonenorm', 'bnorm_epsilon': 0.0001, 'bnorm_running_avg_momentum': 0.001, 'initial_conv_1x1': False, 'trunk_num_channels': 384, 'mid_num_channels': 192, 'gpool_num_channels': 64, 'use_attention_pool': False, 'num_attention_pool_heads': 4, 'block_kind': [['rconv1', 'bottlenest2'], ['rconv2', 'bottlenest2'], ['rconv3', 'bottlenest2gpool'], ['rconv4', 'bottlenest2'], ['rconv5', 'bottlenest2'], ['rconv6', 'bottlenest2gpool'], ['rconv7', 'bottlenest2'], ['rconv8', 'bottlenest2'], ['rconv9', 'bottlenest2gpool'], ['rconv10', 'bottlenest2'], ['rconv11', 'bottlenest2'], ['rconv12', 'bottlenest2gpool'], ['rconv13', 'bottlenest2'], ['rconv14', 'bottlenest2'], ['rconv15', 'bottlenest2gpool'], ['rconv16', 'bottlenest2'], ['rconv17', 'bottlenest2'], ['rconv18', 'bottlenest2']], 'p1_num_channels': 48, 'g1_num_channels': 48, 'v1_num_channels': 96, 'sbv2_num_channels': 112, 'num_scorebeliefs': 8, 'v2_size': 128, 'bnorm_use_gamma': True, 'activation': 'mish', 'use_repvgg_init': True, 'use_repvgg_learning_rate': True, 'has_intermediate_head': True, 'intermediate_head_blocks': 18, 'trunk_normless': True, 'metadata_encoder': {'internal_num_channels': 384}}
|
| 3 |
+
Parameters in model:
|
| 4 |
+
conv_spatial.weight, [384, 22, 3, 3], 76032 params
|
| 5 |
+
linear_global.weight, [384, 19], 7296 params
|
| 6 |
+
metadata_encoder.linear1.weight, [384, 192], 73728 params
|
| 7 |
+
metadata_encoder.linear1.bias, [384], 384 params
|
| 8 |
+
metadata_encoder.linear2.weight, [384, 384], 147456 params
|
| 9 |
+
metadata_encoder.linear2.bias, [384], 384 params
|
| 10 |
+
metadata_encoder.linear_output_to_trunk.weight, [384, 384], 147456 params
|
| 11 |
+
blocks.0.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 12 |
+
blocks.0.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 13 |
+
blocks.0.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 14 |
+
blocks.0.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 15 |
+
blocks.0.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 16 |
+
blocks.0.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 17 |
+
blocks.0.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 18 |
+
blocks.0.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 19 |
+
blocks.0.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 20 |
+
blocks.0.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 21 |
+
blocks.0.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 22 |
+
blocks.0.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 23 |
+
blocks.0.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 24 |
+
blocks.0.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 25 |
+
blocks.0.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 26 |
+
blocks.0.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 27 |
+
blocks.0.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 28 |
+
blocks.0.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 29 |
+
blocks.1.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 30 |
+
blocks.1.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 31 |
+
blocks.1.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 32 |
+
blocks.1.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 33 |
+
blocks.1.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 34 |
+
blocks.1.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 35 |
+
blocks.1.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 36 |
+
blocks.1.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 37 |
+
blocks.1.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 38 |
+
blocks.1.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 39 |
+
blocks.1.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 40 |
+
blocks.1.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 41 |
+
blocks.1.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 42 |
+
blocks.1.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 43 |
+
blocks.1.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 44 |
+
blocks.1.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 45 |
+
blocks.1.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 46 |
+
blocks.1.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 47 |
+
blocks.2.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 48 |
+
blocks.2.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 49 |
+
blocks.2.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 50 |
+
blocks.2.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 51 |
+
blocks.2.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 52 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 53 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 54 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 55 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 56 |
+
blocks.2.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 57 |
+
blocks.2.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 58 |
+
blocks.2.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 59 |
+
blocks.2.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 60 |
+
blocks.2.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 61 |
+
blocks.2.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 62 |
+
blocks.2.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 63 |
+
blocks.2.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 64 |
+
blocks.2.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 65 |
+
blocks.2.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 66 |
+
blocks.2.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 67 |
+
blocks.2.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 68 |
+
blocks.2.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 69 |
+
blocks.3.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 70 |
+
blocks.3.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 71 |
+
blocks.3.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 72 |
+
blocks.3.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 73 |
+
blocks.3.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 74 |
+
blocks.3.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 75 |
+
blocks.3.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 76 |
+
blocks.3.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 77 |
+
blocks.3.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 78 |
+
blocks.3.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 79 |
+
blocks.3.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 80 |
+
blocks.3.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 81 |
+
blocks.3.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 82 |
+
blocks.3.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 83 |
+
blocks.3.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 84 |
+
blocks.3.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 85 |
+
blocks.3.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 86 |
+
blocks.3.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 87 |
+
blocks.4.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 88 |
+
blocks.4.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 89 |
+
blocks.4.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 90 |
+
blocks.4.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 91 |
+
blocks.4.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 92 |
+
blocks.4.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 93 |
+
blocks.4.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 94 |
+
blocks.4.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 95 |
+
blocks.4.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 96 |
+
blocks.4.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 97 |
+
blocks.4.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 98 |
+
blocks.4.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 99 |
+
blocks.4.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 100 |
+
blocks.4.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 101 |
+
blocks.4.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 102 |
+
blocks.4.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 103 |
+
blocks.4.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 104 |
+
blocks.4.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 105 |
+
blocks.5.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 106 |
+
blocks.5.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 107 |
+
blocks.5.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 108 |
+
blocks.5.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 109 |
+
blocks.5.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 110 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 111 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 112 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 113 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 114 |
+
blocks.5.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 115 |
+
blocks.5.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 116 |
+
blocks.5.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 117 |
+
blocks.5.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 118 |
+
blocks.5.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 119 |
+
blocks.5.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 120 |
+
blocks.5.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 121 |
+
blocks.5.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 122 |
+
blocks.5.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 123 |
+
blocks.5.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 124 |
+
blocks.5.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 125 |
+
blocks.5.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 126 |
+
blocks.5.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 127 |
+
blocks.6.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 128 |
+
blocks.6.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 129 |
+
blocks.6.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 130 |
+
blocks.6.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 131 |
+
blocks.6.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 132 |
+
blocks.6.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 133 |
+
blocks.6.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 134 |
+
blocks.6.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 135 |
+
blocks.6.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 136 |
+
blocks.6.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 137 |
+
blocks.6.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 138 |
+
blocks.6.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 139 |
+
blocks.6.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 140 |
+
blocks.6.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 141 |
+
blocks.6.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 142 |
+
blocks.6.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 143 |
+
blocks.6.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 144 |
+
blocks.6.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 145 |
+
blocks.7.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 146 |
+
blocks.7.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 147 |
+
blocks.7.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 148 |
+
blocks.7.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 149 |
+
blocks.7.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 150 |
+
blocks.7.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 151 |
+
blocks.7.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 152 |
+
blocks.7.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 153 |
+
blocks.7.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 154 |
+
blocks.7.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 155 |
+
blocks.7.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 156 |
+
blocks.7.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 157 |
+
blocks.7.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 158 |
+
blocks.7.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 159 |
+
blocks.7.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 160 |
+
blocks.7.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 161 |
+
blocks.7.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 162 |
+
blocks.7.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 163 |
+
blocks.8.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 164 |
+
blocks.8.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 165 |
+
blocks.8.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 166 |
+
blocks.8.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 167 |
+
blocks.8.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 168 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 169 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 170 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 171 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 172 |
+
blocks.8.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 173 |
+
blocks.8.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 174 |
+
blocks.8.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 175 |
+
blocks.8.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 176 |
+
blocks.8.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 177 |
+
blocks.8.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 178 |
+
blocks.8.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 179 |
+
blocks.8.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 180 |
+
blocks.8.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 181 |
+
blocks.8.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 182 |
+
blocks.8.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 183 |
+
blocks.8.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 184 |
+
blocks.8.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 185 |
+
blocks.9.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 186 |
+
blocks.9.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 187 |
+
blocks.9.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 188 |
+
blocks.9.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 189 |
+
blocks.9.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 190 |
+
blocks.9.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 191 |
+
blocks.9.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 192 |
+
blocks.9.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 193 |
+
blocks.9.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 194 |
+
blocks.9.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 195 |
+
blocks.9.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 196 |
+
blocks.9.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 197 |
+
blocks.9.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 198 |
+
blocks.9.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 199 |
+
blocks.9.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 200 |
+
blocks.9.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 201 |
+
blocks.9.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 202 |
+
blocks.9.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 203 |
+
blocks.10.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 204 |
+
blocks.10.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 205 |
+
blocks.10.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 206 |
+
blocks.10.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 207 |
+
blocks.10.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 208 |
+
blocks.10.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 209 |
+
blocks.10.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 210 |
+
blocks.10.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 211 |
+
blocks.10.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 212 |
+
blocks.10.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 213 |
+
blocks.10.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 214 |
+
blocks.10.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 215 |
+
blocks.10.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 216 |
+
blocks.10.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 217 |
+
blocks.10.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 218 |
+
blocks.10.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 219 |
+
blocks.10.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 220 |
+
blocks.10.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 221 |
+
blocks.11.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 222 |
+
blocks.11.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 223 |
+
blocks.11.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 224 |
+
blocks.11.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 225 |
+
blocks.11.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 226 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 227 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 228 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 229 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 230 |
+
blocks.11.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 231 |
+
blocks.11.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 232 |
+
blocks.11.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 233 |
+
blocks.11.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 234 |
+
blocks.11.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 235 |
+
blocks.11.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 236 |
+
blocks.11.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 237 |
+
blocks.11.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 238 |
+
blocks.11.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 239 |
+
blocks.11.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 240 |
+
blocks.11.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 241 |
+
blocks.11.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 242 |
+
blocks.11.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 243 |
+
blocks.12.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 244 |
+
blocks.12.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 245 |
+
blocks.12.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 246 |
+
blocks.12.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 247 |
+
blocks.12.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 248 |
+
blocks.12.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 249 |
+
blocks.12.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 250 |
+
blocks.12.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 251 |
+
blocks.12.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 252 |
+
blocks.12.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 253 |
+
blocks.12.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 254 |
+
blocks.12.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 255 |
+
blocks.12.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 256 |
+
blocks.12.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 257 |
+
blocks.12.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 258 |
+
blocks.12.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 259 |
+
blocks.12.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 260 |
+
blocks.12.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 261 |
+
blocks.13.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 262 |
+
blocks.13.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 263 |
+
blocks.13.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 264 |
+
blocks.13.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 265 |
+
blocks.13.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 266 |
+
blocks.13.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 267 |
+
blocks.13.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 268 |
+
blocks.13.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 269 |
+
blocks.13.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 270 |
+
blocks.13.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 271 |
+
blocks.13.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 272 |
+
blocks.13.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 273 |
+
blocks.13.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 274 |
+
blocks.13.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 275 |
+
blocks.13.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 276 |
+
blocks.13.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 277 |
+
blocks.13.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 278 |
+
blocks.13.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 279 |
+
blocks.14.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 280 |
+
blocks.14.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 281 |
+
blocks.14.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 282 |
+
blocks.14.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 283 |
+
blocks.14.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 284 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 285 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 286 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 287 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 288 |
+
blocks.14.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 289 |
+
blocks.14.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 290 |
+
blocks.14.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 291 |
+
blocks.14.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 292 |
+
blocks.14.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 293 |
+
blocks.14.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 294 |
+
blocks.14.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 295 |
+
blocks.14.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 296 |
+
blocks.14.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 297 |
+
blocks.14.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 298 |
+
blocks.14.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 299 |
+
blocks.14.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 300 |
+
blocks.14.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 301 |
+
blocks.15.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 302 |
+
blocks.15.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 303 |
+
blocks.15.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 304 |
+
blocks.15.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 305 |
+
blocks.15.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 306 |
+
blocks.15.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 307 |
+
blocks.15.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 308 |
+
blocks.15.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 309 |
+
blocks.15.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 310 |
+
blocks.15.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 311 |
+
blocks.15.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 312 |
+
blocks.15.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 313 |
+
blocks.15.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 314 |
+
blocks.15.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 315 |
+
blocks.15.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 316 |
+
blocks.15.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 317 |
+
blocks.15.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 318 |
+
blocks.15.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 319 |
+
blocks.16.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 320 |
+
blocks.16.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 321 |
+
blocks.16.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 322 |
+
blocks.16.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 323 |
+
blocks.16.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 324 |
+
blocks.16.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 325 |
+
blocks.16.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 326 |
+
blocks.16.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 327 |
+
blocks.16.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 328 |
+
blocks.16.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 329 |
+
blocks.16.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 330 |
+
blocks.16.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 331 |
+
blocks.16.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 332 |
+
blocks.16.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 333 |
+
blocks.16.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 334 |
+
blocks.16.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 335 |
+
blocks.16.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 336 |
+
blocks.16.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 337 |
+
blocks.17.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 338 |
+
blocks.17.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 339 |
+
blocks.17.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 340 |
+
blocks.17.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 341 |
+
blocks.17.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 342 |
+
blocks.17.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 343 |
+
blocks.17.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 344 |
+
blocks.17.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 345 |
+
blocks.17.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 346 |
+
blocks.17.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 347 |
+
blocks.17.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 348 |
+
blocks.17.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 349 |
+
blocks.17.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 350 |
+
blocks.17.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 351 |
+
blocks.17.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 352 |
+
blocks.17.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 353 |
+
blocks.17.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 354 |
+
blocks.17.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 355 |
+
norm_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 356 |
+
policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 357 |
+
policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 358 |
+
policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 359 |
+
policy_head.linear_g.weight, [48, 144], 6912 params
|
| 360 |
+
policy_head.linear_pass.weight, [6, 144], 864 params
|
| 361 |
+
policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 362 |
+
policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 363 |
+
value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 364 |
+
value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 365 |
+
value_head.linear2.weight, [128, 288], 36864 params
|
| 366 |
+
value_head.linear2.bias, [128], 128 params
|
| 367 |
+
value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 368 |
+
value_head.linear_valuehead.bias, [3], 3 params
|
| 369 |
+
value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 370 |
+
value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 371 |
+
value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 372 |
+
value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 373 |
+
value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 374 |
+
value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 375 |
+
value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 376 |
+
value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 377 |
+
value_head.linear_s2.weight, [112, 288], 32256 params
|
| 378 |
+
value_head.linear_s2.bias, [112], 112 params
|
| 379 |
+
value_head.linear_s2off.weight, [112, 1], 112 params
|
| 380 |
+
value_head.linear_s2par.weight, [112, 1], 112 params
|
| 381 |
+
value_head.linear_s3.weight, [8, 112], 896 params
|
| 382 |
+
value_head.linear_s3.bias, [8], 8 params
|
| 383 |
+
value_head.linear_smix.weight, [8, 288], 2304 params
|
| 384 |
+
value_head.linear_smix.bias, [8], 8 params
|
| 385 |
+
norm_intermediate_trunkfinal.gamma, [1, 384, 1, 1], 384 params
|
| 386 |
+
norm_intermediate_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 387 |
+
intermediate_policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 388 |
+
intermediate_policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 389 |
+
intermediate_policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 390 |
+
intermediate_policy_head.linear_g.weight, [48, 144], 6912 params
|
| 391 |
+
intermediate_policy_head.linear_pass.weight, [6, 144], 864 params
|
| 392 |
+
intermediate_policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 393 |
+
intermediate_policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 394 |
+
intermediate_value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 395 |
+
intermediate_value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 396 |
+
intermediate_value_head.linear2.weight, [128, 288], 36864 params
|
| 397 |
+
intermediate_value_head.linear2.bias, [128], 128 params
|
| 398 |
+
intermediate_value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 399 |
+
intermediate_value_head.linear_valuehead.bias, [3], 3 params
|
| 400 |
+
intermediate_value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 401 |
+
intermediate_value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 402 |
+
intermediate_value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 403 |
+
intermediate_value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 404 |
+
intermediate_value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 405 |
+
intermediate_value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 406 |
+
intermediate_value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 407 |
+
intermediate_value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 408 |
+
intermediate_value_head.linear_s2.weight, [112, 288], 32256 params
|
| 409 |
+
intermediate_value_head.linear_s2.bias, [112], 112 params
|
| 410 |
+
intermediate_value_head.linear_s2off.weight, [112, 1], 112 params
|
| 411 |
+
intermediate_value_head.linear_s2par.weight, [112, 1], 112 params
|
| 412 |
+
intermediate_value_head.linear_s3.weight, [8, 112], 896 params
|
| 413 |
+
intermediate_value_head.linear_s3.bias, [8], 8 params
|
| 414 |
+
intermediate_value_head.linear_smix.weight, [8, 288], 2304 params
|
| 415 |
+
intermediate_value_head.linear_smix.bias, [8], 8 params
|
| 416 |
+
Total num params: 26934250
|
| 417 |
+
Total trainable params: 26934250
|
| 418 |
+
Writing model
|
| 419 |
+
Exported at:
|
| 420 |
+
2026-06-02 05:54:12.255098 UTC
|
cho-chikun_export/metadata.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"global_step_samples": 10265216, "total_num_data_rows": 542241, "extra_stats": {"sums": {"p0loss_sum": 25148.61060135443, "p1loss_sum": 0.0, "p0softloss_sum": 184949.7611365048, "p1softloss_sum": 0.0, "vloss_sum": 24299.002293833662, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 6.716187737132702, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 23388.20803668426, "pacc1_sum": 23602.45829785597, "vsquare_sum": 12306.545074126434, "wsum": 10265216.0, "nsamp": 10265216.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 175697.7539062353, "sekiweightscale_sum": 255999.99999998524, "norm_input_batch": 4918.519284169386, "norm_input_noreg_batch": 267.9978971574035, "norm_normal_batch": 24230.18304047732, "norm_normal_gamma_batch": 16377.741245632813, "norm_noreg_batch": 8800.381751925981, "norm_output_batch": 7861.214602477442, "norm_output_noreg_batch": 1789.739137298897, "Ip0loss_sum": 30427.815970625878, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 183214.55664519191, "Ip1softloss_sum": 0.0, "Ivloss_sum": 20791.743677197283, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 1.7957252907040713, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 23119.139746343957, "Ivsquare_sum": 8739.446180685629, "pslr_batch": 0.009710010010010007, "wdnormal_batch": 0.01773150979609149, "gnorm_cap_batch": 242750.25025025025, "window_start_batch": 0.0, "window_end_batch": 52651655.37837838, "gnorm_batch": 27728.22201309994, "exgnorm_sum": 0.0}, "weights": {"p0loss_sum": 31999.999999998156, "p1loss_sum": 31999.999999998156, "p0softloss_sum": 31999.999999998156, "p1softloss_sum": 31999.999999998156, "vloss_sum": 31999.999999998156, "tdvloss1_sum": 31999.999999998156, "tdvloss2_sum": 31999.999999998156, "tdvloss3_sum": 31999.999999998156, "tdsloss_sum": 31999.999999998156, "oloss_sum": 31999.999999998156, "sloss_sum": 31999.999999998156, "fploss_sum": 31999.999999998156, "skloss_sum": 31999.999999998156, "smloss_sum": 31999.999999998156, "sbcdfloss_sum": 31999.999999998156, "sbpdfloss_sum": 31999.999999998156, "sdregloss_sum": 31999.999999998156, "leadloss_sum": 31999.999999998156, "vtimeloss_sum": 31999.999999998156, "evstloss_sum": 31999.999999998156, "esstloss_sum": 31999.999999998156, "qwlloss_sum": 31999.999999998156, "qscloss_sum": 31999.999999998156, "loss_sum": 31999.999999998156, "pacc1_sum": 31999.999999998156, "vsquare_sum": 31999.999999998156, "wsum": 10265216.0, "nsamp": 10265216.0, "ptentr_sum": 31999.999999998156, "ptsoftentr_sum": 31999.999999998156, "sekiweightscale_sum": 31999.999999998156, "norm_input_batch": 97.1001001001001, "norm_input_noreg_batch": 97.1001001001001, "norm_normal_batch": 97.1001001001001, "norm_normal_gamma_batch": 97.1001001001001, "norm_noreg_batch": 97.1001001001001, "norm_output_batch": 97.1001001001001, "norm_output_noreg_batch": 97.1001001001001, "Ip0loss_sum": 31999.999999998156, "Ip1loss_sum": 31999.999999998156, "Ip0softloss_sum": 31999.999999998156, "Ip1softloss_sum": 31999.999999998156, "Ivloss_sum": 31999.999999998156, "Itdvloss1_sum": 31999.999999998156, "Itdvloss2_sum": 31999.999999998156, "Itdvloss3_sum": 31999.999999998156, "Itdsloss_sum": 31999.999999998156, "Ioloss_sum": 31999.999999998156, "Isloss_sum": 31999.999999998156, "Ifploss_sum": 31999.999999998156, "Iskloss_sum": 31999.999999998156, "Ismloss_sum": 31999.999999998156, "Isbcdfloss_sum": 31999.999999998156, "Isbpdfloss_sum": 31999.999999998156, "Isdregloss_sum": 31999.999999998156, "Ileadloss_sum": 31999.999999998156, "Ivtimeloss_sum": 31999.999999998156, "Ievstloss_sum": 31999.999999998156, "Iesstloss_sum": 31999.999999998156, "Iqwlloss_sum": 31999.999999998156, "Iqscloss_sum": 31999.999999998156, "Ipacc1_sum": 31999.999999998156, "Ivsquare_sum": 31999.999999998156, "pslr_batch": 97.1001001001001, "wdnormal_batch": 97.1001001001001, "gnorm_cap_batch": 97.1001001001001, "window_start_batch": 97.1001001001001, "window_end_batch": 97.1001001001001, "gnorm_batch": 97.1001001001001, "exgnorm_sum": 31984.831240176543}, "last_val_metrics": {"sums": {"p0loss_sum": 1453.45210647583, "p1loss_sum": 0.0, "p0softloss_sum": 5018.563323974609, "p1softloss_sum": 0.0, "vloss_sum": 794.46375413239, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 0.15692153736017644, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 1636.4731674194336, "pacc1_sum": 490.0, "vsquare_sum": 354.2902886867523, "wsum": 864.0, "nsamp": 864.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 4743.83935546875, "sekiweightscale_sum": 6048.0, "norm_input_batch": 1.36769588470459, "norm_input_noreg_batch": 0.07445500874519348, "norm_normal_batch": 6.734721862792969, "norm_normal_gamma_batch": 4.554408828735352, "norm_noreg_batch": 2.44697428894043, "norm_output_batch": 2.1858538513183596, "norm_output_noreg_batch": 0.49764755058288573, "Ip0loss_sum": 1297.271183013916, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 4975.360794067383, "Ip1softloss_sum": 0.0, "Ivloss_sum": 636.5729194879532, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 0.049053263530367985, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 501.0, "Ivsquare_sum": 249.83517980575562, "nsamp_train": 10230112.0, "wsum_train": 10230112.0}, "weights": {"p0loss_sum": 864.0, "p1loss_sum": 864.0, "p0softloss_sum": 864.0, "p1softloss_sum": 864.0, "vloss_sum": 864.0, "tdvloss1_sum": 864.0, "tdvloss2_sum": 864.0, "tdvloss3_sum": 864.0, "tdsloss_sum": 864.0, "oloss_sum": 864.0, "sloss_sum": 864.0, "fploss_sum": 864.0, "skloss_sum": 864.0, "smloss_sum": 864.0, "sbcdfloss_sum": 864.0, "sbpdfloss_sum": 864.0, "sdregloss_sum": 864.0, "leadloss_sum": 864.0, "vtimeloss_sum": 864.0, "evstloss_sum": 864.0, "esstloss_sum": 864.0, "qwlloss_sum": 864.0, "qscloss_sum": 864.0, "loss_sum": 864.0, "pacc1_sum": 864.0, "vsquare_sum": 864.0, "wsum": 864.0, "nsamp": 864.0, "ptentr_sum": 864.0, "ptsoftentr_sum": 864.0, "sekiweightscale_sum": 864.0, "norm_input_batch": 0.027, "norm_input_noreg_batch": 0.027, "norm_normal_batch": 0.027, "norm_normal_gamma_batch": 0.027, "norm_noreg_batch": 0.027, "norm_output_batch": 0.027, "norm_output_noreg_batch": 0.027, "Ip0loss_sum": 864.0, "Ip1loss_sum": 864.0, "Ip0softloss_sum": 864.0, "Ip1softloss_sum": 864.0, "Ivloss_sum": 864.0, "Itdvloss1_sum": 864.0, "Itdvloss2_sum": 864.0, "Itdvloss3_sum": 864.0, "Itdsloss_sum": 864.0, "Ioloss_sum": 864.0, "Isloss_sum": 864.0, "Ifploss_sum": 864.0, "Iskloss_sum": 864.0, "Ismloss_sum": 864.0, "Isbcdfloss_sum": 864.0, "Isbpdfloss_sum": 864.0, "Isdregloss_sum": 864.0, "Ileadloss_sum": 864.0, "Ivtimeloss_sum": 864.0, "Ievstloss_sum": 864.0, "Iesstloss_sum": 864.0, "Iqwlloss_sum": 864.0, "Iqscloss_sum": 864.0, "Ipacc1_sum": 864.0, "Ivsquare_sum": 864.0, "nsamp_train": 10230112.0, "wsum_train": 10230112.0}}}}
|
cho-chikun_export/model.bin.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1032f9c3779a59004636502ea783553f99a2840615c1313783f7cc8e0df84ec
|
| 3 |
+
size 99050761
|
kitani_export/best_checkpoint.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9870898a1d5f1cab1caa23875b569ad68a3bb4f6fb6dccd1dc5ad78a454f4d6
|
| 3 |
+
size 214677331
|
kitani_export/log.txt
ADDED
|
@@ -0,0 +1,420 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
['/home/ran/go-arm-rl/KataGo.OLD/python/export_model_pytorch.py', '-checkpoint', '/home/ran/go-arm-rl/models/kitani_75/best_checkpoint.ckpt', '-export-dir', '/home/ran/go-arm-rl/models/kitani_75', '-model-name', 'kitani_75', '-filename-prefix', 'model', '-export-14-as-15']
|
| 2 |
+
{'version': 14, 'norm_kind': 'fixscaleonenorm', 'bnorm_epsilon': 0.0001, 'bnorm_running_avg_momentum': 0.001, 'initial_conv_1x1': False, 'trunk_num_channels': 384, 'mid_num_channels': 192, 'gpool_num_channels': 64, 'use_attention_pool': False, 'num_attention_pool_heads': 4, 'block_kind': [['rconv1', 'bottlenest2'], ['rconv2', 'bottlenest2'], ['rconv3', 'bottlenest2gpool'], ['rconv4', 'bottlenest2'], ['rconv5', 'bottlenest2'], ['rconv6', 'bottlenest2gpool'], ['rconv7', 'bottlenest2'], ['rconv8', 'bottlenest2'], ['rconv9', 'bottlenest2gpool'], ['rconv10', 'bottlenest2'], ['rconv11', 'bottlenest2'], ['rconv12', 'bottlenest2gpool'], ['rconv13', 'bottlenest2'], ['rconv14', 'bottlenest2'], ['rconv15', 'bottlenest2gpool'], ['rconv16', 'bottlenest2'], ['rconv17', 'bottlenest2'], ['rconv18', 'bottlenest2']], 'p1_num_channels': 48, 'g1_num_channels': 48, 'v1_num_channels': 96, 'sbv2_num_channels': 112, 'num_scorebeliefs': 8, 'v2_size': 128, 'bnorm_use_gamma': True, 'activation': 'mish', 'use_repvgg_init': True, 'use_repvgg_learning_rate': True, 'has_intermediate_head': True, 'intermediate_head_blocks': 18, 'trunk_normless': True, 'metadata_encoder': {'internal_num_channels': 384}}
|
| 3 |
+
Parameters in model:
|
| 4 |
+
conv_spatial.weight, [384, 22, 3, 3], 76032 params
|
| 5 |
+
linear_global.weight, [384, 19], 7296 params
|
| 6 |
+
metadata_encoder.linear1.weight, [384, 192], 73728 params
|
| 7 |
+
metadata_encoder.linear1.bias, [384], 384 params
|
| 8 |
+
metadata_encoder.linear2.weight, [384, 384], 147456 params
|
| 9 |
+
metadata_encoder.linear2.bias, [384], 384 params
|
| 10 |
+
metadata_encoder.linear_output_to_trunk.weight, [384, 384], 147456 params
|
| 11 |
+
blocks.0.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 12 |
+
blocks.0.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 13 |
+
blocks.0.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 14 |
+
blocks.0.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 15 |
+
blocks.0.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 16 |
+
blocks.0.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 17 |
+
blocks.0.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 18 |
+
blocks.0.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 19 |
+
blocks.0.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 20 |
+
blocks.0.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 21 |
+
blocks.0.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 22 |
+
blocks.0.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 23 |
+
blocks.0.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 24 |
+
blocks.0.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 25 |
+
blocks.0.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 26 |
+
blocks.0.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 27 |
+
blocks.0.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 28 |
+
blocks.0.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 29 |
+
blocks.1.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 30 |
+
blocks.1.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 31 |
+
blocks.1.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 32 |
+
blocks.1.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 33 |
+
blocks.1.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 34 |
+
blocks.1.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 35 |
+
blocks.1.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 36 |
+
blocks.1.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 37 |
+
blocks.1.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 38 |
+
blocks.1.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 39 |
+
blocks.1.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 40 |
+
blocks.1.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 41 |
+
blocks.1.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 42 |
+
blocks.1.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 43 |
+
blocks.1.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 44 |
+
blocks.1.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 45 |
+
blocks.1.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 46 |
+
blocks.1.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 47 |
+
blocks.2.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 48 |
+
blocks.2.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 49 |
+
blocks.2.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 50 |
+
blocks.2.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 51 |
+
blocks.2.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 52 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 53 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 54 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 55 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 56 |
+
blocks.2.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 57 |
+
blocks.2.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 58 |
+
blocks.2.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 59 |
+
blocks.2.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 60 |
+
blocks.2.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 61 |
+
blocks.2.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 62 |
+
blocks.2.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 63 |
+
blocks.2.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 64 |
+
blocks.2.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 65 |
+
blocks.2.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 66 |
+
blocks.2.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 67 |
+
blocks.2.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 68 |
+
blocks.2.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 69 |
+
blocks.3.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 70 |
+
blocks.3.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 71 |
+
blocks.3.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 72 |
+
blocks.3.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 73 |
+
blocks.3.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 74 |
+
blocks.3.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 75 |
+
blocks.3.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 76 |
+
blocks.3.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 77 |
+
blocks.3.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 78 |
+
blocks.3.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 79 |
+
blocks.3.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 80 |
+
blocks.3.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 81 |
+
blocks.3.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 82 |
+
blocks.3.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 83 |
+
blocks.3.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 84 |
+
blocks.3.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 85 |
+
blocks.3.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 86 |
+
blocks.3.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 87 |
+
blocks.4.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 88 |
+
blocks.4.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 89 |
+
blocks.4.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 90 |
+
blocks.4.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 91 |
+
blocks.4.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 92 |
+
blocks.4.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 93 |
+
blocks.4.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 94 |
+
blocks.4.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 95 |
+
blocks.4.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 96 |
+
blocks.4.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 97 |
+
blocks.4.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 98 |
+
blocks.4.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 99 |
+
blocks.4.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 100 |
+
blocks.4.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 101 |
+
blocks.4.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 102 |
+
blocks.4.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 103 |
+
blocks.4.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 104 |
+
blocks.4.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 105 |
+
blocks.5.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 106 |
+
blocks.5.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 107 |
+
blocks.5.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 108 |
+
blocks.5.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 109 |
+
blocks.5.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 110 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 111 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 112 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 113 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 114 |
+
blocks.5.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 115 |
+
blocks.5.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 116 |
+
blocks.5.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 117 |
+
blocks.5.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 118 |
+
blocks.5.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 119 |
+
blocks.5.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 120 |
+
blocks.5.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 121 |
+
blocks.5.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 122 |
+
blocks.5.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 123 |
+
blocks.5.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 124 |
+
blocks.5.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 125 |
+
blocks.5.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 126 |
+
blocks.5.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 127 |
+
blocks.6.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 128 |
+
blocks.6.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 129 |
+
blocks.6.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 130 |
+
blocks.6.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 131 |
+
blocks.6.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 132 |
+
blocks.6.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 133 |
+
blocks.6.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 134 |
+
blocks.6.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 135 |
+
blocks.6.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 136 |
+
blocks.6.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 137 |
+
blocks.6.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 138 |
+
blocks.6.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 139 |
+
blocks.6.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 140 |
+
blocks.6.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 141 |
+
blocks.6.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 142 |
+
blocks.6.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 143 |
+
blocks.6.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 144 |
+
blocks.6.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 145 |
+
blocks.7.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 146 |
+
blocks.7.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 147 |
+
blocks.7.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 148 |
+
blocks.7.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 149 |
+
blocks.7.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 150 |
+
blocks.7.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 151 |
+
blocks.7.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 152 |
+
blocks.7.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 153 |
+
blocks.7.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 154 |
+
blocks.7.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 155 |
+
blocks.7.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 156 |
+
blocks.7.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 157 |
+
blocks.7.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 158 |
+
blocks.7.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 159 |
+
blocks.7.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 160 |
+
blocks.7.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 161 |
+
blocks.7.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 162 |
+
blocks.7.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 163 |
+
blocks.8.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 164 |
+
blocks.8.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 165 |
+
blocks.8.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 166 |
+
blocks.8.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 167 |
+
blocks.8.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 168 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 169 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 170 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 171 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 172 |
+
blocks.8.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 173 |
+
blocks.8.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 174 |
+
blocks.8.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 175 |
+
blocks.8.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 176 |
+
blocks.8.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 177 |
+
blocks.8.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 178 |
+
blocks.8.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 179 |
+
blocks.8.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 180 |
+
blocks.8.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 181 |
+
blocks.8.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 182 |
+
blocks.8.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 183 |
+
blocks.8.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 184 |
+
blocks.8.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 185 |
+
blocks.9.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 186 |
+
blocks.9.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 187 |
+
blocks.9.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 188 |
+
blocks.9.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 189 |
+
blocks.9.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 190 |
+
blocks.9.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 191 |
+
blocks.9.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 192 |
+
blocks.9.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 193 |
+
blocks.9.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 194 |
+
blocks.9.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 195 |
+
blocks.9.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 196 |
+
blocks.9.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 197 |
+
blocks.9.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 198 |
+
blocks.9.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 199 |
+
blocks.9.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 200 |
+
blocks.9.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 201 |
+
blocks.9.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 202 |
+
blocks.9.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 203 |
+
blocks.10.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 204 |
+
blocks.10.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 205 |
+
blocks.10.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 206 |
+
blocks.10.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 207 |
+
blocks.10.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 208 |
+
blocks.10.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 209 |
+
blocks.10.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 210 |
+
blocks.10.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 211 |
+
blocks.10.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 212 |
+
blocks.10.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 213 |
+
blocks.10.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 214 |
+
blocks.10.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 215 |
+
blocks.10.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 216 |
+
blocks.10.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 217 |
+
blocks.10.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 218 |
+
blocks.10.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 219 |
+
blocks.10.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 220 |
+
blocks.10.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 221 |
+
blocks.11.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 222 |
+
blocks.11.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 223 |
+
blocks.11.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 224 |
+
blocks.11.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 225 |
+
blocks.11.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 226 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 227 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 228 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 229 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 230 |
+
blocks.11.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 231 |
+
blocks.11.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 232 |
+
blocks.11.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 233 |
+
blocks.11.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 234 |
+
blocks.11.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 235 |
+
blocks.11.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 236 |
+
blocks.11.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 237 |
+
blocks.11.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 238 |
+
blocks.11.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 239 |
+
blocks.11.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 240 |
+
blocks.11.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 241 |
+
blocks.11.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 242 |
+
blocks.11.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 243 |
+
blocks.12.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 244 |
+
blocks.12.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 245 |
+
blocks.12.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 246 |
+
blocks.12.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 247 |
+
blocks.12.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 248 |
+
blocks.12.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 249 |
+
blocks.12.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 250 |
+
blocks.12.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 251 |
+
blocks.12.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 252 |
+
blocks.12.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 253 |
+
blocks.12.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 254 |
+
blocks.12.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 255 |
+
blocks.12.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 256 |
+
blocks.12.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 257 |
+
blocks.12.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 258 |
+
blocks.12.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 259 |
+
blocks.12.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 260 |
+
blocks.12.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 261 |
+
blocks.13.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 262 |
+
blocks.13.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 263 |
+
blocks.13.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 264 |
+
blocks.13.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 265 |
+
blocks.13.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 266 |
+
blocks.13.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 267 |
+
blocks.13.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 268 |
+
blocks.13.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 269 |
+
blocks.13.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 270 |
+
blocks.13.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 271 |
+
blocks.13.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 272 |
+
blocks.13.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 273 |
+
blocks.13.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 274 |
+
blocks.13.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 275 |
+
blocks.13.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 276 |
+
blocks.13.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 277 |
+
blocks.13.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 278 |
+
blocks.13.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 279 |
+
blocks.14.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 280 |
+
blocks.14.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 281 |
+
blocks.14.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 282 |
+
blocks.14.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 283 |
+
blocks.14.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 284 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 285 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 286 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 287 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 288 |
+
blocks.14.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 289 |
+
blocks.14.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 290 |
+
blocks.14.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 291 |
+
blocks.14.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 292 |
+
blocks.14.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 293 |
+
blocks.14.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 294 |
+
blocks.14.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 295 |
+
blocks.14.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 296 |
+
blocks.14.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 297 |
+
blocks.14.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 298 |
+
blocks.14.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 299 |
+
blocks.14.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 300 |
+
blocks.14.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 301 |
+
blocks.15.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 302 |
+
blocks.15.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 303 |
+
blocks.15.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 304 |
+
blocks.15.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 305 |
+
blocks.15.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 306 |
+
blocks.15.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 307 |
+
blocks.15.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 308 |
+
blocks.15.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 309 |
+
blocks.15.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 310 |
+
blocks.15.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 311 |
+
blocks.15.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 312 |
+
blocks.15.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 313 |
+
blocks.15.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 314 |
+
blocks.15.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 315 |
+
blocks.15.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 316 |
+
blocks.15.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 317 |
+
blocks.15.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 318 |
+
blocks.15.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 319 |
+
blocks.16.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 320 |
+
blocks.16.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 321 |
+
blocks.16.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 322 |
+
blocks.16.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 323 |
+
blocks.16.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 324 |
+
blocks.16.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 325 |
+
blocks.16.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 326 |
+
blocks.16.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 327 |
+
blocks.16.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 328 |
+
blocks.16.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 329 |
+
blocks.16.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 330 |
+
blocks.16.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 331 |
+
blocks.16.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 332 |
+
blocks.16.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 333 |
+
blocks.16.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 334 |
+
blocks.16.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 335 |
+
blocks.16.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 336 |
+
blocks.16.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 337 |
+
blocks.17.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 338 |
+
blocks.17.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 339 |
+
blocks.17.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 340 |
+
blocks.17.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 341 |
+
blocks.17.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 342 |
+
blocks.17.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 343 |
+
blocks.17.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 344 |
+
blocks.17.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 345 |
+
blocks.17.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 346 |
+
blocks.17.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 347 |
+
blocks.17.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 348 |
+
blocks.17.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 349 |
+
blocks.17.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 350 |
+
blocks.17.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 351 |
+
blocks.17.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 352 |
+
blocks.17.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 353 |
+
blocks.17.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 354 |
+
blocks.17.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 355 |
+
norm_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 356 |
+
policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 357 |
+
policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 358 |
+
policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 359 |
+
policy_head.linear_g.weight, [48, 144], 6912 params
|
| 360 |
+
policy_head.linear_pass.weight, [6, 144], 864 params
|
| 361 |
+
policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 362 |
+
policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 363 |
+
value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 364 |
+
value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 365 |
+
value_head.linear2.weight, [128, 288], 36864 params
|
| 366 |
+
value_head.linear2.bias, [128], 128 params
|
| 367 |
+
value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 368 |
+
value_head.linear_valuehead.bias, [3], 3 params
|
| 369 |
+
value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 370 |
+
value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 371 |
+
value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 372 |
+
value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 373 |
+
value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 374 |
+
value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 375 |
+
value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 376 |
+
value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 377 |
+
value_head.linear_s2.weight, [112, 288], 32256 params
|
| 378 |
+
value_head.linear_s2.bias, [112], 112 params
|
| 379 |
+
value_head.linear_s2off.weight, [112, 1], 112 params
|
| 380 |
+
value_head.linear_s2par.weight, [112, 1], 112 params
|
| 381 |
+
value_head.linear_s3.weight, [8, 112], 896 params
|
| 382 |
+
value_head.linear_s3.bias, [8], 8 params
|
| 383 |
+
value_head.linear_smix.weight, [8, 288], 2304 params
|
| 384 |
+
value_head.linear_smix.bias, [8], 8 params
|
| 385 |
+
norm_intermediate_trunkfinal.gamma, [1, 384, 1, 1], 384 params
|
| 386 |
+
norm_intermediate_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 387 |
+
intermediate_policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 388 |
+
intermediate_policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 389 |
+
intermediate_policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 390 |
+
intermediate_policy_head.linear_g.weight, [48, 144], 6912 params
|
| 391 |
+
intermediate_policy_head.linear_pass.weight, [6, 144], 864 params
|
| 392 |
+
intermediate_policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 393 |
+
intermediate_policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 394 |
+
intermediate_value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 395 |
+
intermediate_value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 396 |
+
intermediate_value_head.linear2.weight, [128, 288], 36864 params
|
| 397 |
+
intermediate_value_head.linear2.bias, [128], 128 params
|
| 398 |
+
intermediate_value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 399 |
+
intermediate_value_head.linear_valuehead.bias, [3], 3 params
|
| 400 |
+
intermediate_value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 401 |
+
intermediate_value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 402 |
+
intermediate_value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 403 |
+
intermediate_value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 404 |
+
intermediate_value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 405 |
+
intermediate_value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 406 |
+
intermediate_value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 407 |
+
intermediate_value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 408 |
+
intermediate_value_head.linear_s2.weight, [112, 288], 32256 params
|
| 409 |
+
intermediate_value_head.linear_s2.bias, [112], 112 params
|
| 410 |
+
intermediate_value_head.linear_s2off.weight, [112, 1], 112 params
|
| 411 |
+
intermediate_value_head.linear_s2par.weight, [112, 1], 112 params
|
| 412 |
+
intermediate_value_head.linear_s3.weight, [8, 112], 896 params
|
| 413 |
+
intermediate_value_head.linear_s3.bias, [8], 8 params
|
| 414 |
+
intermediate_value_head.linear_smix.weight, [8, 288], 2304 params
|
| 415 |
+
intermediate_value_head.linear_smix.bias, [8], 8 params
|
| 416 |
+
Total num params: 26934250
|
| 417 |
+
Total trainable params: 26934250
|
| 418 |
+
Writing model
|
| 419 |
+
Exported at:
|
| 420 |
+
2026-06-02 05:54:23.596133 UTC
|
kitani_export/metadata.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"global_step_samples": 2682240, "total_num_data_rows": 223541, "extra_stats": {"sums": {"p0loss_sum": 25959.507236803587, "p1loss_sum": 0.0, "p0softloss_sum": 181477.1466342303, "p1softloss_sum": 0.0, "vloss_sum": 24541.861193241148, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 4.475904426579813, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 24142.341907692837, "pacc1_sum": 23660.676059689154, "vsquare_sum": 12771.512890175813, "wsum": 2682240.0, "nsamp": 2682240.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 175697.7539062353, "sekiweightscale_sum": 255999.99999998524, "norm_input_batch": 3345.23115073307, "norm_input_noreg_batch": 176.27287525865282, "norm_normal_batch": 15043.785312915095, "norm_normal_gamma_batch": 10905.363859998386, "norm_noreg_batch": 5730.206188957848, "norm_output_batch": 5182.221117757353, "norm_output_noreg_batch": 1181.8616241361192, "Ip0loss_sum": 30667.200796386638, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 183787.69542442917, "Ip1softloss_sum": 0.0, "Ivloss_sum": 21653.046597763016, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 2.7365379932653187, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 23139.405895706463, "Ivsquare_sum": 10538.943828828551, "pslr_batch": 0.006410010010010014, "wdnormal_batch": 0.010314352199478793, "gnorm_cap_batch": 160250.25025025025, "window_start_batch": 0.0, "window_end_batch": 14329000.476476476, "gnorm_batch": 18383.5870724671, "exgnorm_sum": 0.0}, "weights": {"p0loss_sum": 31999.999999998156, "p1loss_sum": 31999.999999998156, "p0softloss_sum": 31999.999999998156, "p1softloss_sum": 31999.999999998156, "vloss_sum": 31999.999999998156, "tdvloss1_sum": 31999.999999998156, "tdvloss2_sum": 31999.999999998156, "tdvloss3_sum": 31999.999999998156, "tdsloss_sum": 31999.999999998156, "oloss_sum": 31999.999999998156, "sloss_sum": 31999.999999998156, "fploss_sum": 31999.999999998156, "skloss_sum": 31999.999999998156, "smloss_sum": 31999.999999998156, "sbcdfloss_sum": 31999.999999998156, "sbpdfloss_sum": 31999.999999998156, "sdregloss_sum": 31999.999999998156, "leadloss_sum": 31999.999999998156, "vtimeloss_sum": 31999.999999998156, "evstloss_sum": 31999.999999998156, "esstloss_sum": 31999.999999998156, "qwlloss_sum": 31999.999999998156, "qscloss_sum": 31999.999999998156, "loss_sum": 31999.999999998156, "pacc1_sum": 31999.999999998156, "vsquare_sum": 31999.999999998156, "wsum": 2682240.0, "nsamp": 2682240.0, "ptentr_sum": 31999.999999998156, "ptsoftentr_sum": 31999.999999998156, "sekiweightscale_sum": 31999.999999998156, "norm_input_batch": 64.1001001001001, "norm_input_noreg_batch": 64.1001001001001, "norm_normal_batch": 64.1001001001001, "norm_normal_gamma_batch": 64.1001001001001, "norm_noreg_batch": 64.1001001001001, "norm_output_batch": 64.1001001001001, "norm_output_noreg_batch": 64.1001001001001, "Ip0loss_sum": 31999.999999998156, "Ip1loss_sum": 31999.999999998156, "Ip0softloss_sum": 31999.999999998156, "Ip1softloss_sum": 31999.999999998156, "Ivloss_sum": 31999.999999998156, "Itdvloss1_sum": 31999.999999998156, "Itdvloss2_sum": 31999.999999998156, "Itdvloss3_sum": 31999.999999998156, "Itdsloss_sum": 31999.999999998156, "Ioloss_sum": 31999.999999998156, "Isloss_sum": 31999.999999998156, "Ifploss_sum": 31999.999999998156, "Iskloss_sum": 31999.999999998156, "Ismloss_sum": 31999.999999998156, "Isbcdfloss_sum": 31999.999999998156, "Isbpdfloss_sum": 31999.999999998156, "Isdregloss_sum": 31999.999999998156, "Ileadloss_sum": 31999.999999998156, "Ivtimeloss_sum": 31999.999999998156, "Ievstloss_sum": 31999.999999998156, "Iesstloss_sum": 31999.999999998156, "Iqwlloss_sum": 31999.999999998156, "Iqscloss_sum": 31999.999999998156, "Ipacc1_sum": 31999.999999998156, "Ivsquare_sum": 31999.999999998156, "pslr_batch": 64.1001001001001, "wdnormal_batch": 64.1001001001001, "gnorm_cap_batch": 64.1001001001001, "window_start_batch": 64.1001001001001, "window_end_batch": 64.1001001001001, "gnorm_batch": 64.1001001001001, "exgnorm_sum": 31984.76313307108}, "last_val_metrics": {"sums": {"p0loss_sum": 2053.6502685546875, "p1loss_sum": 0.0, "p0softloss_sum": 6583.764862060547, "p1softloss_sum": 0.0, "vloss_sum": 1195.8098721951246, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 0.13394801681715762, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 2291.257022857666, "pacc1_sum": 628.0, "vsquare_sum": 533.4917516708374, "wsum": 1152.0, "nsamp": 1152.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 6325.119140625, "sekiweightscale_sum": 8064.0, "norm_input_batch": 1.8791496276855468, "norm_input_noreg_batch": 0.09903230667114259, "norm_normal_batch": 8.437531311035157, "norm_normal_gamma_batch": 6.124519226074219, "norm_noreg_batch": 3.2174464416503907, "norm_output_batch": 2.91031787109375, "norm_output_noreg_batch": 0.6637644195556641, "Ip0loss_sum": 1718.636079788208, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 6666.4927978515625, "Ip1softloss_sum": 0.0, "Ivloss_sum": 1014.1555427610874, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 0.3676987288054079, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 655.0, "Ivsquare_sum": 445.5287902355194, "nsamp_train": 2609792.0, "wsum_train": 2609792.0}, "weights": {"p0loss_sum": 1152.0, "p1loss_sum": 1152.0, "p0softloss_sum": 1152.0, "p1softloss_sum": 1152.0, "vloss_sum": 1152.0, "tdvloss1_sum": 1152.0, "tdvloss2_sum": 1152.0, "tdvloss3_sum": 1152.0, "tdsloss_sum": 1152.0, "oloss_sum": 1152.0, "sloss_sum": 1152.0, "fploss_sum": 1152.0, "skloss_sum": 1152.0, "smloss_sum": 1152.0, "sbcdfloss_sum": 1152.0, "sbpdfloss_sum": 1152.0, "sdregloss_sum": 1152.0, "leadloss_sum": 1152.0, "vtimeloss_sum": 1152.0, "evstloss_sum": 1152.0, "esstloss_sum": 1152.0, "qwlloss_sum": 1152.0, "qscloss_sum": 1152.0, "loss_sum": 1152.0, "pacc1_sum": 1152.0, "vsquare_sum": 1152.0, "wsum": 1152.0, "nsamp": 1152.0, "ptentr_sum": 1152.0, "ptsoftentr_sum": 1152.0, "sekiweightscale_sum": 1152.0, "norm_input_batch": 0.036000000000000004, "norm_input_noreg_batch": 0.036000000000000004, "norm_normal_batch": 0.036000000000000004, "norm_normal_gamma_batch": 0.036000000000000004, "norm_noreg_batch": 0.036000000000000004, "norm_output_batch": 0.036000000000000004, "norm_output_noreg_batch": 0.036000000000000004, "Ip0loss_sum": 1152.0, "Ip1loss_sum": 1152.0, "Ip0softloss_sum": 1152.0, "Ip1softloss_sum": 1152.0, "Ivloss_sum": 1152.0, "Itdvloss1_sum": 1152.0, "Itdvloss2_sum": 1152.0, "Itdvloss3_sum": 1152.0, "Itdsloss_sum": 1152.0, "Ioloss_sum": 1152.0, "Isloss_sum": 1152.0, "Ifploss_sum": 1152.0, "Iskloss_sum": 1152.0, "Ismloss_sum": 1152.0, "Isbcdfloss_sum": 1152.0, "Isbpdfloss_sum": 1152.0, "Isdregloss_sum": 1152.0, "Ileadloss_sum": 1152.0, "Ivtimeloss_sum": 1152.0, "Ievstloss_sum": 1152.0, "Iesstloss_sum": 1152.0, "Iqwlloss_sum": 1152.0, "Iqscloss_sum": 1152.0, "Ipacc1_sum": 1152.0, "Ivsquare_sum": 1152.0, "nsamp_train": 2609792.0, "wsum_train": 2609792.0}}}}
|
kitani_export/model.bin.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9cc54c927d1462558d1a034a32141843a53d720df83ecd969fb161205e83a66f
|
| 3 |
+
size 99067070
|
nieweiping_export/best_checkpoint.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d9aca1c48b35fc5f98cbefb68bc28f0127fb70e2b91d55e754b17eb4b86f4435
|
| 3 |
+
size 230600501
|
nieweiping_export/log.txt
ADDED
|
@@ -0,0 +1,482 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
['/home/ran/go-arm-rl/KataGo.OLD/python/export_model_pytorch.py', '-checkpoint', '/home/ran/go-arm-rl/models/nieweiping_75/best_checkpoint.ckpt', '-export-dir', '/home/ran/go-arm-rl/models/nieweiping_75', '-model-name', 'nieweiping_75', '-filename-prefix', 'model']
|
| 2 |
+
{'version': 14, 'norm_kind': 'fixscaleonenorm', 'bnorm_epsilon': 0.0001, 'bnorm_running_avg_momentum': 0.001, 'initial_conv_1x1': False, 'trunk_num_channels': 384, 'mid_num_channels': 192, 'gpool_num_channels': 64, 'use_attention_pool': False, 'num_attention_pool_heads': 4, 'block_kind': [['rconv1', 'bottlenest2'], ['rconv2', 'bottlenest2'], ['rconv3', 'bottlenest2gpool'], ['rconv4', 'bottlenest2'], ['rconv5', 'bottlenest2'], ['rconv6', 'bottlenest2gpool'], ['rconv7', 'bottlenest2'], ['rconv8', 'bottlenest2'], ['rconv9', 'bottlenest2gpool'], ['rconv10', 'bottlenest2'], ['rconv11', 'bottlenest2'], ['rconv12', 'bottlenest2gpool'], ['rconv13', 'bottlenest2'], ['rconv14', 'bottlenest2'], ['rconv15', 'bottlenest2gpool'], ['rconv16', 'bottlenest2'], ['rconv17', 'bottlenest2'], ['rconv18', 'bottlenest2']], 'p1_num_channels': 48, 'g1_num_channels': 48, 'v1_num_channels': 48, 'sbv2_num_channels': 112, 'num_scorebeliefs': 8, 'v2_size': 128, 'bnorm_use_gamma': True, 'activation': 'mish', 'use_repvgg_linear': True, 'has_intermediate_head': True, 'intermediate_head_blocks': 18, 'trunk_normless': True}
|
| 3 |
+
Parameters in model:
|
| 4 |
+
conv_spatial.weight, [384, 22, 3, 3], 76032 params
|
| 5 |
+
linear_global.weight, [384, 19], 7296 params
|
| 6 |
+
blocks.0.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 7 |
+
blocks.0.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 8 |
+
blocks.0.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 9 |
+
blocks.0.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 10 |
+
blocks.0.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 11 |
+
blocks.0.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 12 |
+
blocks.0.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 13 |
+
blocks.0.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 14 |
+
blocks.0.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 15 |
+
blocks.0.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 16 |
+
blocks.0.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 17 |
+
blocks.0.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 18 |
+
blocks.0.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 19 |
+
blocks.0.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 20 |
+
blocks.0.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 21 |
+
blocks.0.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 22 |
+
blocks.0.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 23 |
+
blocks.0.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 24 |
+
blocks.0.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 25 |
+
blocks.0.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 26 |
+
blocks.0.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 27 |
+
blocks.0.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 28 |
+
blocks.1.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 29 |
+
blocks.1.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 30 |
+
blocks.1.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 31 |
+
blocks.1.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 32 |
+
blocks.1.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 33 |
+
blocks.1.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 34 |
+
blocks.1.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 35 |
+
blocks.1.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 36 |
+
blocks.1.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 37 |
+
blocks.1.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 38 |
+
blocks.1.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 39 |
+
blocks.1.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 40 |
+
blocks.1.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 41 |
+
blocks.1.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 42 |
+
blocks.1.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 43 |
+
blocks.1.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 44 |
+
blocks.1.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 45 |
+
blocks.1.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 46 |
+
blocks.1.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 47 |
+
blocks.1.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 48 |
+
blocks.1.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 49 |
+
blocks.1.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 50 |
+
blocks.2.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 51 |
+
blocks.2.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 52 |
+
blocks.2.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 53 |
+
blocks.2.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 54 |
+
blocks.2.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 55 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 56 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 57 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 58 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 59 |
+
blocks.2.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 60 |
+
blocks.2.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 61 |
+
blocks.2.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 62 |
+
blocks.2.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 63 |
+
blocks.2.blockstack.0.normactconv2.conv1x1.weight, [192, 128, 1, 1], 24576 params
|
| 64 |
+
blocks.2.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 65 |
+
blocks.2.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 66 |
+
blocks.2.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 67 |
+
blocks.2.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 68 |
+
blocks.2.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 69 |
+
blocks.2.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 70 |
+
blocks.2.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 71 |
+
blocks.2.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 72 |
+
blocks.2.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 73 |
+
blocks.2.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 74 |
+
blocks.2.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 75 |
+
blocks.3.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 76 |
+
blocks.3.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 77 |
+
blocks.3.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 78 |
+
blocks.3.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 79 |
+
blocks.3.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 80 |
+
blocks.3.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 81 |
+
blocks.3.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 82 |
+
blocks.3.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 83 |
+
blocks.3.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 84 |
+
blocks.3.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 85 |
+
blocks.3.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 86 |
+
blocks.3.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 87 |
+
blocks.3.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 88 |
+
blocks.3.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 89 |
+
blocks.3.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 90 |
+
blocks.3.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 91 |
+
blocks.3.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 92 |
+
blocks.3.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 93 |
+
blocks.3.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 94 |
+
blocks.3.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 95 |
+
blocks.3.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 96 |
+
blocks.3.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 97 |
+
blocks.4.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 98 |
+
blocks.4.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 99 |
+
blocks.4.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 100 |
+
blocks.4.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 101 |
+
blocks.4.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 102 |
+
blocks.4.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 103 |
+
blocks.4.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 104 |
+
blocks.4.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 105 |
+
blocks.4.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 106 |
+
blocks.4.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 107 |
+
blocks.4.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 108 |
+
blocks.4.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 109 |
+
blocks.4.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 110 |
+
blocks.4.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 111 |
+
blocks.4.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 112 |
+
blocks.4.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 113 |
+
blocks.4.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 114 |
+
blocks.4.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 115 |
+
blocks.4.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 116 |
+
blocks.4.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 117 |
+
blocks.4.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 118 |
+
blocks.4.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 119 |
+
blocks.5.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 120 |
+
blocks.5.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 121 |
+
blocks.5.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 122 |
+
blocks.5.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 123 |
+
blocks.5.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 124 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 125 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 126 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 127 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 128 |
+
blocks.5.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 129 |
+
blocks.5.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 130 |
+
blocks.5.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 131 |
+
blocks.5.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 132 |
+
blocks.5.blockstack.0.normactconv2.conv1x1.weight, [192, 128, 1, 1], 24576 params
|
| 133 |
+
blocks.5.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 134 |
+
blocks.5.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 135 |
+
blocks.5.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 136 |
+
blocks.5.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 137 |
+
blocks.5.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 138 |
+
blocks.5.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 139 |
+
blocks.5.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 140 |
+
blocks.5.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 141 |
+
blocks.5.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 142 |
+
blocks.5.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 143 |
+
blocks.5.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 144 |
+
blocks.6.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 145 |
+
blocks.6.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 146 |
+
blocks.6.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 147 |
+
blocks.6.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 148 |
+
blocks.6.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 149 |
+
blocks.6.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 150 |
+
blocks.6.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 151 |
+
blocks.6.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 152 |
+
blocks.6.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 153 |
+
blocks.6.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 154 |
+
blocks.6.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 155 |
+
blocks.6.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 156 |
+
blocks.6.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 157 |
+
blocks.6.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 158 |
+
blocks.6.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 159 |
+
blocks.6.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 160 |
+
blocks.6.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 161 |
+
blocks.6.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 162 |
+
blocks.6.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 163 |
+
blocks.6.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 164 |
+
blocks.6.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 165 |
+
blocks.6.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 166 |
+
blocks.7.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 167 |
+
blocks.7.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 168 |
+
blocks.7.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 169 |
+
blocks.7.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 170 |
+
blocks.7.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 171 |
+
blocks.7.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 172 |
+
blocks.7.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 173 |
+
blocks.7.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 174 |
+
blocks.7.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 175 |
+
blocks.7.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 176 |
+
blocks.7.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 177 |
+
blocks.7.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 178 |
+
blocks.7.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 179 |
+
blocks.7.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 180 |
+
blocks.7.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 181 |
+
blocks.7.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 182 |
+
blocks.7.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 183 |
+
blocks.7.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 184 |
+
blocks.7.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 185 |
+
blocks.7.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 186 |
+
blocks.7.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 187 |
+
blocks.7.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 188 |
+
blocks.8.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 189 |
+
blocks.8.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 190 |
+
blocks.8.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 191 |
+
blocks.8.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 192 |
+
blocks.8.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 193 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 194 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 195 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 196 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 197 |
+
blocks.8.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 198 |
+
blocks.8.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 199 |
+
blocks.8.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 200 |
+
blocks.8.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 201 |
+
blocks.8.blockstack.0.normactconv2.conv1x1.weight, [192, 128, 1, 1], 24576 params
|
| 202 |
+
blocks.8.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 203 |
+
blocks.8.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 204 |
+
blocks.8.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 205 |
+
blocks.8.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 206 |
+
blocks.8.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 207 |
+
blocks.8.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 208 |
+
blocks.8.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 209 |
+
blocks.8.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 210 |
+
blocks.8.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 211 |
+
blocks.8.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 212 |
+
blocks.8.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 213 |
+
blocks.9.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 214 |
+
blocks.9.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 215 |
+
blocks.9.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 216 |
+
blocks.9.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 217 |
+
blocks.9.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 218 |
+
blocks.9.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 219 |
+
blocks.9.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 220 |
+
blocks.9.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 221 |
+
blocks.9.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 222 |
+
blocks.9.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 223 |
+
blocks.9.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 224 |
+
blocks.9.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 225 |
+
blocks.9.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 226 |
+
blocks.9.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 227 |
+
blocks.9.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 228 |
+
blocks.9.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 229 |
+
blocks.9.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 230 |
+
blocks.9.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 231 |
+
blocks.9.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 232 |
+
blocks.9.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 233 |
+
blocks.9.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 234 |
+
blocks.9.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 235 |
+
blocks.10.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 236 |
+
blocks.10.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 237 |
+
blocks.10.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 238 |
+
blocks.10.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 239 |
+
blocks.10.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 240 |
+
blocks.10.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 241 |
+
blocks.10.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 242 |
+
blocks.10.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 243 |
+
blocks.10.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 244 |
+
blocks.10.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 245 |
+
blocks.10.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 246 |
+
blocks.10.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 247 |
+
blocks.10.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 248 |
+
blocks.10.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 249 |
+
blocks.10.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 250 |
+
blocks.10.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 251 |
+
blocks.10.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 252 |
+
blocks.10.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 253 |
+
blocks.10.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 254 |
+
blocks.10.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 255 |
+
blocks.10.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 256 |
+
blocks.10.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 257 |
+
blocks.11.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 258 |
+
blocks.11.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 259 |
+
blocks.11.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 260 |
+
blocks.11.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 261 |
+
blocks.11.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 262 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 263 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 264 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 265 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 266 |
+
blocks.11.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 267 |
+
blocks.11.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 268 |
+
blocks.11.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 269 |
+
blocks.11.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 270 |
+
blocks.11.blockstack.0.normactconv2.conv1x1.weight, [192, 128, 1, 1], 24576 params
|
| 271 |
+
blocks.11.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 272 |
+
blocks.11.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 273 |
+
blocks.11.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 274 |
+
blocks.11.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 275 |
+
blocks.11.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 276 |
+
blocks.11.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 277 |
+
blocks.11.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 278 |
+
blocks.11.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 279 |
+
blocks.11.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 280 |
+
blocks.11.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 281 |
+
blocks.11.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 282 |
+
blocks.12.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 283 |
+
blocks.12.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 284 |
+
blocks.12.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 285 |
+
blocks.12.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 286 |
+
blocks.12.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 287 |
+
blocks.12.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 288 |
+
blocks.12.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 289 |
+
blocks.12.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 290 |
+
blocks.12.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 291 |
+
blocks.12.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 292 |
+
blocks.12.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 293 |
+
blocks.12.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 294 |
+
blocks.12.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 295 |
+
blocks.12.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 296 |
+
blocks.12.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 297 |
+
blocks.12.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 298 |
+
blocks.12.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 299 |
+
blocks.12.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 300 |
+
blocks.12.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 301 |
+
blocks.12.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 302 |
+
blocks.12.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 303 |
+
blocks.12.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 304 |
+
blocks.13.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 305 |
+
blocks.13.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 306 |
+
blocks.13.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 307 |
+
blocks.13.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 308 |
+
blocks.13.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 309 |
+
blocks.13.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 310 |
+
blocks.13.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 311 |
+
blocks.13.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 312 |
+
blocks.13.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 313 |
+
blocks.13.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 314 |
+
blocks.13.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 315 |
+
blocks.13.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 316 |
+
blocks.13.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 317 |
+
blocks.13.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 318 |
+
blocks.13.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 319 |
+
blocks.13.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 320 |
+
blocks.13.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 321 |
+
blocks.13.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 322 |
+
blocks.13.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 323 |
+
blocks.13.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 324 |
+
blocks.13.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 325 |
+
blocks.13.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 326 |
+
blocks.14.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 327 |
+
blocks.14.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 328 |
+
blocks.14.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 329 |
+
blocks.14.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 330 |
+
blocks.14.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 331 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 332 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 333 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 334 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 335 |
+
blocks.14.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 336 |
+
blocks.14.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 337 |
+
blocks.14.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 338 |
+
blocks.14.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 339 |
+
blocks.14.blockstack.0.normactconv2.conv1x1.weight, [192, 128, 1, 1], 24576 params
|
| 340 |
+
blocks.14.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 341 |
+
blocks.14.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 342 |
+
blocks.14.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 343 |
+
blocks.14.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 344 |
+
blocks.14.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 345 |
+
blocks.14.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 346 |
+
blocks.14.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 347 |
+
blocks.14.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 348 |
+
blocks.14.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 349 |
+
blocks.14.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 350 |
+
blocks.14.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 351 |
+
blocks.15.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 352 |
+
blocks.15.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 353 |
+
blocks.15.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 354 |
+
blocks.15.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 355 |
+
blocks.15.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 356 |
+
blocks.15.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 357 |
+
blocks.15.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 358 |
+
blocks.15.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 359 |
+
blocks.15.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 360 |
+
blocks.15.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 361 |
+
blocks.15.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 362 |
+
blocks.15.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 363 |
+
blocks.15.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 364 |
+
blocks.15.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 365 |
+
blocks.15.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 366 |
+
blocks.15.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 367 |
+
blocks.15.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 368 |
+
blocks.15.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 369 |
+
blocks.15.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 370 |
+
blocks.15.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 371 |
+
blocks.15.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 372 |
+
blocks.15.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 373 |
+
blocks.16.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 374 |
+
blocks.16.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 375 |
+
blocks.16.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 376 |
+
blocks.16.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 377 |
+
blocks.16.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 378 |
+
blocks.16.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 379 |
+
blocks.16.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 380 |
+
blocks.16.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 381 |
+
blocks.16.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 382 |
+
blocks.16.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 383 |
+
blocks.16.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 384 |
+
blocks.16.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 385 |
+
blocks.16.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 386 |
+
blocks.16.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 387 |
+
blocks.16.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 388 |
+
blocks.16.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 389 |
+
blocks.16.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 390 |
+
blocks.16.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 391 |
+
blocks.16.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 392 |
+
blocks.16.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 393 |
+
blocks.16.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 394 |
+
blocks.16.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 395 |
+
blocks.17.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 396 |
+
blocks.17.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 397 |
+
blocks.17.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 398 |
+
blocks.17.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 399 |
+
blocks.17.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 400 |
+
blocks.17.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 401 |
+
blocks.17.blockstack.0.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 402 |
+
blocks.17.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 403 |
+
blocks.17.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 404 |
+
blocks.17.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 405 |
+
blocks.17.blockstack.0.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 406 |
+
blocks.17.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 407 |
+
blocks.17.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 408 |
+
blocks.17.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 409 |
+
blocks.17.blockstack.1.normactconv1.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 410 |
+
blocks.17.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 411 |
+
blocks.17.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 412 |
+
blocks.17.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 413 |
+
blocks.17.blockstack.1.normactconv2.conv1x1.weight, [192, 192, 1, 1], 36864 params
|
| 414 |
+
blocks.17.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 415 |
+
blocks.17.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 416 |
+
blocks.17.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 417 |
+
norm_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 418 |
+
policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 419 |
+
policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 420 |
+
policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 421 |
+
policy_head.linear_g.weight, [48, 144], 6912 params
|
| 422 |
+
policy_head.linear_pass.weight, [6, 144], 864 params
|
| 423 |
+
policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 424 |
+
policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 425 |
+
value_head.conv1.weight, [48, 384, 1, 1], 18432 params
|
| 426 |
+
value_head.bias1.beta, [1, 48, 1, 1], 48 params
|
| 427 |
+
value_head.linear2.weight, [128, 144], 18432 params
|
| 428 |
+
value_head.linear2.bias, [128], 128 params
|
| 429 |
+
value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 430 |
+
value_head.linear_valuehead.bias, [3], 3 params
|
| 431 |
+
value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 432 |
+
value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 433 |
+
value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 434 |
+
value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 435 |
+
value_head.conv_ownership.weight, [1, 48, 1, 1], 48 params
|
| 436 |
+
value_head.conv_scoring.weight, [1, 48, 1, 1], 48 params
|
| 437 |
+
value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 438 |
+
value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 439 |
+
value_head.linear_s2.weight, [112, 144], 16128 params
|
| 440 |
+
value_head.linear_s2.bias, [112], 112 params
|
| 441 |
+
value_head.linear_s2off.weight, [112, 1], 112 params
|
| 442 |
+
value_head.linear_s2par.weight, [112, 1], 112 params
|
| 443 |
+
value_head.linear_s3.weight, [8, 112], 896 params
|
| 444 |
+
value_head.linear_s3.bias, [8], 8 params
|
| 445 |
+
value_head.linear_smix.weight, [8, 144], 1152 params
|
| 446 |
+
value_head.linear_smix.bias, [8], 8 params
|
| 447 |
+
norm_intermediate_trunkfinal.gamma, [1, 384, 1, 1], 384 params
|
| 448 |
+
norm_intermediate_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 449 |
+
intermediate_policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 450 |
+
intermediate_policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 451 |
+
intermediate_policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 452 |
+
intermediate_policy_head.linear_g.weight, [48, 144], 6912 params
|
| 453 |
+
intermediate_policy_head.linear_pass.weight, [6, 144], 864 params
|
| 454 |
+
intermediate_policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 455 |
+
intermediate_policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 456 |
+
intermediate_value_head.conv1.weight, [48, 384, 1, 1], 18432 params
|
| 457 |
+
intermediate_value_head.bias1.beta, [1, 48, 1, 1], 48 params
|
| 458 |
+
intermediate_value_head.linear2.weight, [128, 144], 18432 params
|
| 459 |
+
intermediate_value_head.linear2.bias, [128], 128 params
|
| 460 |
+
intermediate_value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 461 |
+
intermediate_value_head.linear_valuehead.bias, [3], 3 params
|
| 462 |
+
intermediate_value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 463 |
+
intermediate_value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 464 |
+
intermediate_value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 465 |
+
intermediate_value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 466 |
+
intermediate_value_head.conv_ownership.weight, [1, 48, 1, 1], 48 params
|
| 467 |
+
intermediate_value_head.conv_scoring.weight, [1, 48, 1, 1], 48 params
|
| 468 |
+
intermediate_value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 469 |
+
intermediate_value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 470 |
+
intermediate_value_head.linear_s2.weight, [112, 144], 16128 params
|
| 471 |
+
intermediate_value_head.linear_s2.bias, [112], 112 params
|
| 472 |
+
intermediate_value_head.linear_s2off.weight, [112, 1], 112 params
|
| 473 |
+
intermediate_value_head.linear_s2par.weight, [112, 1], 112 params
|
| 474 |
+
intermediate_value_head.linear_s3.weight, [8, 112], 896 params
|
| 475 |
+
intermediate_value_head.linear_s3.bias, [8], 8 params
|
| 476 |
+
intermediate_value_head.linear_smix.weight, [8, 144], 1152 params
|
| 477 |
+
intermediate_value_head.linear_smix.bias, [8], 8 params
|
| 478 |
+
Total num params: 28864714
|
| 479 |
+
Total trainable params: 28864714
|
| 480 |
+
Writing model
|
| 481 |
+
Exported at:
|
| 482 |
+
2026-06-02 05:53:24.792357 UTC
|
nieweiping_export/metadata.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"global_step_samples": 3404544, "total_num_data_rows": 208289, "extra_stats": {"sums": {"p0loss_sum": 42629.08307439693, "p1loss_sum": 0.0, "p0softloss_sum": 379946.5917360415, "p1softloss_sum": 0.0, "vloss_sum": 109966.14351099127, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 310.528329369887, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 39645.04755777488, "pacc1_sum": 49879.28207388182, "vsquare_sum": 47415.77051841554, "wsum": 3404544.0, "nsamp": 3404544.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 351395.5078124706, "sekiweightscale_sum": 511999.9999999705, "norm_input_batch": 866.9385249425768, "norm_input_noreg_batch": 0.0, "norm_normal_batch": 8494.281006468904, "norm_normal_gamma_batch": 6309.451527625698, "norm_noreg_batch": 5577.355289945343, "norm_output_batch": 4001.650637736179, "norm_output_noreg_batch": 1131.2294007816533, "Ip0loss_sum": 63453.618848578146, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 368055.875786236, "Ip1softloss_sum": 0.0, "Ivloss_sum": 42497.753894372625, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 12.196344367420473, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 48650.8123022683, "Ivsquare_sum": 27880.484727264957, "pslr_batch": 0.0032100100100100086, "wdnormal_batch": 0.013211448953392976, "gnorm_cap_batch": 80250.25025025025, "window_start_batch": 0.0, "window_end_batch": 6686097.74974975, "gnorm_batch": 14523.969257350647, "exgnorm_sum": 0.0}, "weights": {"p0loss_sum": 63999.99999999631, "p1loss_sum": 63999.99999999631, "p0softloss_sum": 63999.99999999631, "p1softloss_sum": 63999.99999999631, "vloss_sum": 63999.99999999631, "tdvloss1_sum": 63999.99999999631, "tdvloss2_sum": 63999.99999999631, "tdvloss3_sum": 63999.99999999631, "tdsloss_sum": 63999.99999999631, "oloss_sum": 63999.99999999631, "sloss_sum": 63999.99999999631, "fploss_sum": 63999.99999999631, "skloss_sum": 63999.99999999631, "smloss_sum": 63999.99999999631, "sbcdfloss_sum": 63999.99999999631, "sbpdfloss_sum": 63999.99999999631, "sdregloss_sum": 63999.99999999631, "leadloss_sum": 63999.99999999631, "vtimeloss_sum": 63999.99999999631, "evstloss_sum": 63999.99999999631, "esstloss_sum": 63999.99999999631, "qwlloss_sum": 63999.99999999631, "qscloss_sum": 63999.99999999631, "loss_sum": 63999.99999999631, "pacc1_sum": 63999.99999999631, "vsquare_sum": 63999.99999999631, "wsum": 3404544.0, "nsamp": 3404544.0, "ptentr_sum": 63999.99999999631, "ptsoftentr_sum": 63999.99999999631, "sekiweightscale_sum": 63999.99999999631, "norm_input_batch": 32.1001001001001, "norm_input_noreg_batch": 32.1001001001001, "norm_normal_batch": 32.1001001001001, "norm_normal_gamma_batch": 32.1001001001001, "norm_noreg_batch": 32.1001001001001, "norm_output_batch": 32.1001001001001, "norm_output_noreg_batch": 32.1001001001001, "Ip0loss_sum": 63999.99999999631, "Ip1loss_sum": 63999.99999999631, "Ip0softloss_sum": 63999.99999999631, "Ip1softloss_sum": 63999.99999999631, "Ivloss_sum": 63999.99999999631, "Itdvloss1_sum": 63999.99999999631, "Itdvloss2_sum": 63999.99999999631, "Itdvloss3_sum": 63999.99999999631, "Itdsloss_sum": 63999.99999999631, "Ioloss_sum": 63999.99999999631, "Isloss_sum": 63999.99999999631, "Ifploss_sum": 63999.99999999631, "Iskloss_sum": 63999.99999999631, "Ismloss_sum": 63999.99999999631, "Isbcdfloss_sum": 63999.99999999631, "Isbpdfloss_sum": 63999.99999999631, "Isdregloss_sum": 63999.99999999631, "Ileadloss_sum": 63999.99999999631, "Ivtimeloss_sum": 63999.99999999631, "Ievstloss_sum": 63999.99999999631, "Iesstloss_sum": 63999.99999999631, "Iqwlloss_sum": 63999.99999999631, "Iqscloss_sum": 63999.99999999631, "Ipacc1_sum": 63999.99999999631, "Ivsquare_sum": 63999.99999999631, "pslr_batch": 32.1001001001001, "wdnormal_batch": 32.1001001001001, "gnorm_cap_batch": 32.1001001001001, "window_start_batch": 32.1001001001001, "window_end_batch": 32.1001001001001, "gnorm_batch": 32.1001001001001, "exgnorm_sum": 63994.41563332977}, "last_val_metrics": {"sums": {"p0loss_sum": 2294.880546569824, "p1loss_sum": 0.0, "p0softloss_sum": 7315.641052246094, "p1softloss_sum": 0.0, "vloss_sum": 2154.4328168183565, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 5.884664369747043, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 2626.454635620117, "pacc1_sum": 616.0, "vsquare_sum": 902.0334339141846, "wsum": 1216.0, "nsamp": 1216.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 6676.5146484375, "sekiweightscale_sum": 8512.0, "norm_input_batch": 0.5128076400756836, "norm_input_noreg_batch": 0.0, "norm_normal_batch": 5.022571258544922, "norm_normal_gamma_batch": 3.735446212768555, "norm_noreg_batch": 3.300884994506836, "norm_output_batch": 2.3685196304321288, "norm_output_noreg_batch": 0.669570686340332, "Ip0loss_sum": 2128.860221862793, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 7061.069488525391, "Ip1softloss_sum": 0.0, "Ivloss_sum": 741.5231919288635, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 0.11844930215738714, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 604.0, "Ivsquare_sum": 617.8747634887695, "nsamp_train": 3332096.0, "wsum_train": 3332096.0}, "weights": {"p0loss_sum": 1216.0, "p1loss_sum": 1216.0, "p0softloss_sum": 1216.0, "p1softloss_sum": 1216.0, "vloss_sum": 1216.0, "tdvloss1_sum": 1216.0, "tdvloss2_sum": 1216.0, "tdvloss3_sum": 1216.0, "tdsloss_sum": 1216.0, "oloss_sum": 1216.0, "sloss_sum": 1216.0, "fploss_sum": 1216.0, "skloss_sum": 1216.0, "smloss_sum": 1216.0, "sbcdfloss_sum": 1216.0, "sbpdfloss_sum": 1216.0, "sdregloss_sum": 1216.0, "leadloss_sum": 1216.0, "vtimeloss_sum": 1216.0, "evstloss_sum": 1216.0, "esstloss_sum": 1216.0, "qwlloss_sum": 1216.0, "qscloss_sum": 1216.0, "loss_sum": 1216.0, "pacc1_sum": 1216.0, "vsquare_sum": 1216.0, "wsum": 1216.0, "nsamp": 1216.0, "ptentr_sum": 1216.0, "ptsoftentr_sum": 1216.0, "sekiweightscale_sum": 1216.0, "norm_input_batch": 0.019, "norm_input_noreg_batch": 0.019, "norm_normal_batch": 0.019, "norm_normal_gamma_batch": 0.019, "norm_noreg_batch": 0.019, "norm_output_batch": 0.019, "norm_output_noreg_batch": 0.019, "Ip0loss_sum": 1216.0, "Ip1loss_sum": 1216.0, "Ip0softloss_sum": 1216.0, "Ip1softloss_sum": 1216.0, "Ivloss_sum": 1216.0, "Itdvloss1_sum": 1216.0, "Itdvloss2_sum": 1216.0, "Itdvloss3_sum": 1216.0, "Itdsloss_sum": 1216.0, "Ioloss_sum": 1216.0, "Isloss_sum": 1216.0, "Ifploss_sum": 1216.0, "Iskloss_sum": 1216.0, "Ismloss_sum": 1216.0, "Isbcdfloss_sum": 1216.0, "Isbpdfloss_sum": 1216.0, "Isdregloss_sum": 1216.0, "Ileadloss_sum": 1216.0, "Ivtimeloss_sum": 1216.0, "Ievstloss_sum": 1216.0, "Iesstloss_sum": 1216.0, "Iqwlloss_sum": 1216.0, "Iqscloss_sum": 1216.0, "Ipacc1_sum": 1216.0, "Ivsquare_sum": 1216.0, "nsamp_train": 3332096.0, "wsum_train": 3332096.0}}}}
|
nieweiping_export/model.bin.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:073c5819df351eeecc34f3effd718e88553e8b4281ea3f747654faeba5187be2
|
| 3 |
+
size 97601596
|
shusaku_export/best_checkpoint.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:297f7b5cab9784253055d35cea12f45515186dab39f84071f0ee8cd3a42a0a46
|
| 3 |
+
size 214677139
|
shusaku_export/log.txt
ADDED
|
@@ -0,0 +1,420 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
['/home/ran/go-arm-rl/KataGo.OLD/python/export_model_pytorch.py', '-checkpoint', '/home/ran/go-arm-rl/models/shusaku_75/best_checkpoint.ckpt', '-export-dir', '/home/ran/go-arm-rl/models/shusaku_75', '-model-name', 'shusaku_75', '-filename-prefix', 'model', '-export-14-as-15']
|
| 2 |
+
{'version': 14, 'norm_kind': 'fixscaleonenorm', 'bnorm_epsilon': 0.0001, 'bnorm_running_avg_momentum': 0.001, 'initial_conv_1x1': False, 'trunk_num_channels': 384, 'mid_num_channels': 192, 'gpool_num_channels': 64, 'use_attention_pool': False, 'num_attention_pool_heads': 4, 'block_kind': [['rconv1', 'bottlenest2'], ['rconv2', 'bottlenest2'], ['rconv3', 'bottlenest2gpool'], ['rconv4', 'bottlenest2'], ['rconv5', 'bottlenest2'], ['rconv6', 'bottlenest2gpool'], ['rconv7', 'bottlenest2'], ['rconv8', 'bottlenest2'], ['rconv9', 'bottlenest2gpool'], ['rconv10', 'bottlenest2'], ['rconv11', 'bottlenest2'], ['rconv12', 'bottlenest2gpool'], ['rconv13', 'bottlenest2'], ['rconv14', 'bottlenest2'], ['rconv15', 'bottlenest2gpool'], ['rconv16', 'bottlenest2'], ['rconv17', 'bottlenest2'], ['rconv18', 'bottlenest2']], 'p1_num_channels': 48, 'g1_num_channels': 48, 'v1_num_channels': 96, 'sbv2_num_channels': 112, 'num_scorebeliefs': 8, 'v2_size': 128, 'bnorm_use_gamma': True, 'activation': 'mish', 'use_repvgg_init': True, 'use_repvgg_learning_rate': True, 'has_intermediate_head': True, 'intermediate_head_blocks': 18, 'trunk_normless': True, 'metadata_encoder': {'internal_num_channels': 384}}
|
| 3 |
+
Parameters in model:
|
| 4 |
+
conv_spatial.weight, [384, 22, 3, 3], 76032 params
|
| 5 |
+
linear_global.weight, [384, 19], 7296 params
|
| 6 |
+
metadata_encoder.linear1.weight, [384, 192], 73728 params
|
| 7 |
+
metadata_encoder.linear1.bias, [384], 384 params
|
| 8 |
+
metadata_encoder.linear2.weight, [384, 384], 147456 params
|
| 9 |
+
metadata_encoder.linear2.bias, [384], 384 params
|
| 10 |
+
metadata_encoder.linear_output_to_trunk.weight, [384, 384], 147456 params
|
| 11 |
+
blocks.0.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 12 |
+
blocks.0.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 13 |
+
blocks.0.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 14 |
+
blocks.0.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 15 |
+
blocks.0.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 16 |
+
blocks.0.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 17 |
+
blocks.0.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 18 |
+
blocks.0.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 19 |
+
blocks.0.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 20 |
+
blocks.0.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 21 |
+
blocks.0.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 22 |
+
blocks.0.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 23 |
+
blocks.0.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 24 |
+
blocks.0.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 25 |
+
blocks.0.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 26 |
+
blocks.0.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 27 |
+
blocks.0.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 28 |
+
blocks.0.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 29 |
+
blocks.1.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 30 |
+
blocks.1.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 31 |
+
blocks.1.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 32 |
+
blocks.1.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 33 |
+
blocks.1.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 34 |
+
blocks.1.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 35 |
+
blocks.1.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 36 |
+
blocks.1.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 37 |
+
blocks.1.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 38 |
+
blocks.1.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 39 |
+
blocks.1.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 40 |
+
blocks.1.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 41 |
+
blocks.1.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 42 |
+
blocks.1.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 43 |
+
blocks.1.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 44 |
+
blocks.1.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 45 |
+
blocks.1.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 46 |
+
blocks.1.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 47 |
+
blocks.2.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 48 |
+
blocks.2.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 49 |
+
blocks.2.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 50 |
+
blocks.2.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 51 |
+
blocks.2.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 52 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 53 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 54 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 55 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 56 |
+
blocks.2.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 57 |
+
blocks.2.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 58 |
+
blocks.2.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 59 |
+
blocks.2.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 60 |
+
blocks.2.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 61 |
+
blocks.2.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 62 |
+
blocks.2.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 63 |
+
blocks.2.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 64 |
+
blocks.2.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 65 |
+
blocks.2.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 66 |
+
blocks.2.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 67 |
+
blocks.2.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 68 |
+
blocks.2.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 69 |
+
blocks.3.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 70 |
+
blocks.3.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 71 |
+
blocks.3.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 72 |
+
blocks.3.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 73 |
+
blocks.3.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 74 |
+
blocks.3.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 75 |
+
blocks.3.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 76 |
+
blocks.3.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 77 |
+
blocks.3.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 78 |
+
blocks.3.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 79 |
+
blocks.3.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 80 |
+
blocks.3.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 81 |
+
blocks.3.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 82 |
+
blocks.3.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 83 |
+
blocks.3.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 84 |
+
blocks.3.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 85 |
+
blocks.3.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 86 |
+
blocks.3.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 87 |
+
blocks.4.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 88 |
+
blocks.4.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 89 |
+
blocks.4.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 90 |
+
blocks.4.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 91 |
+
blocks.4.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 92 |
+
blocks.4.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 93 |
+
blocks.4.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 94 |
+
blocks.4.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 95 |
+
blocks.4.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 96 |
+
blocks.4.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 97 |
+
blocks.4.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 98 |
+
blocks.4.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 99 |
+
blocks.4.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 100 |
+
blocks.4.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 101 |
+
blocks.4.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 102 |
+
blocks.4.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 103 |
+
blocks.4.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 104 |
+
blocks.4.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 105 |
+
blocks.5.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 106 |
+
blocks.5.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 107 |
+
blocks.5.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 108 |
+
blocks.5.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 109 |
+
blocks.5.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 110 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 111 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 112 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 113 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 114 |
+
blocks.5.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 115 |
+
blocks.5.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 116 |
+
blocks.5.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 117 |
+
blocks.5.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 118 |
+
blocks.5.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 119 |
+
blocks.5.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 120 |
+
blocks.5.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 121 |
+
blocks.5.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 122 |
+
blocks.5.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 123 |
+
blocks.5.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 124 |
+
blocks.5.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 125 |
+
blocks.5.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 126 |
+
blocks.5.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 127 |
+
blocks.6.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 128 |
+
blocks.6.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 129 |
+
blocks.6.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 130 |
+
blocks.6.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 131 |
+
blocks.6.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 132 |
+
blocks.6.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 133 |
+
blocks.6.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 134 |
+
blocks.6.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 135 |
+
blocks.6.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 136 |
+
blocks.6.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 137 |
+
blocks.6.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 138 |
+
blocks.6.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 139 |
+
blocks.6.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 140 |
+
blocks.6.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 141 |
+
blocks.6.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 142 |
+
blocks.6.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 143 |
+
blocks.6.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 144 |
+
blocks.6.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 145 |
+
blocks.7.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 146 |
+
blocks.7.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 147 |
+
blocks.7.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 148 |
+
blocks.7.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 149 |
+
blocks.7.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 150 |
+
blocks.7.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 151 |
+
blocks.7.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 152 |
+
blocks.7.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 153 |
+
blocks.7.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 154 |
+
blocks.7.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 155 |
+
blocks.7.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 156 |
+
blocks.7.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 157 |
+
blocks.7.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 158 |
+
blocks.7.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 159 |
+
blocks.7.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 160 |
+
blocks.7.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 161 |
+
blocks.7.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 162 |
+
blocks.7.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 163 |
+
blocks.8.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 164 |
+
blocks.8.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 165 |
+
blocks.8.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 166 |
+
blocks.8.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 167 |
+
blocks.8.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 168 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 169 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 170 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 171 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 172 |
+
blocks.8.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 173 |
+
blocks.8.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 174 |
+
blocks.8.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 175 |
+
blocks.8.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 176 |
+
blocks.8.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 177 |
+
blocks.8.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 178 |
+
blocks.8.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 179 |
+
blocks.8.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 180 |
+
blocks.8.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 181 |
+
blocks.8.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 182 |
+
blocks.8.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 183 |
+
blocks.8.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 184 |
+
blocks.8.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 185 |
+
blocks.9.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 186 |
+
blocks.9.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 187 |
+
blocks.9.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 188 |
+
blocks.9.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 189 |
+
blocks.9.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 190 |
+
blocks.9.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 191 |
+
blocks.9.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 192 |
+
blocks.9.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 193 |
+
blocks.9.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 194 |
+
blocks.9.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 195 |
+
blocks.9.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 196 |
+
blocks.9.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 197 |
+
blocks.9.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 198 |
+
blocks.9.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 199 |
+
blocks.9.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 200 |
+
blocks.9.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 201 |
+
blocks.9.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 202 |
+
blocks.9.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 203 |
+
blocks.10.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 204 |
+
blocks.10.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 205 |
+
blocks.10.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 206 |
+
blocks.10.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 207 |
+
blocks.10.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 208 |
+
blocks.10.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 209 |
+
blocks.10.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 210 |
+
blocks.10.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 211 |
+
blocks.10.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 212 |
+
blocks.10.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 213 |
+
blocks.10.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 214 |
+
blocks.10.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 215 |
+
blocks.10.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 216 |
+
blocks.10.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 217 |
+
blocks.10.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 218 |
+
blocks.10.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 219 |
+
blocks.10.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 220 |
+
blocks.10.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 221 |
+
blocks.11.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 222 |
+
blocks.11.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 223 |
+
blocks.11.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 224 |
+
blocks.11.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 225 |
+
blocks.11.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 226 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 227 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 228 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 229 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 230 |
+
blocks.11.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 231 |
+
blocks.11.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 232 |
+
blocks.11.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 233 |
+
blocks.11.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 234 |
+
blocks.11.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 235 |
+
blocks.11.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 236 |
+
blocks.11.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 237 |
+
blocks.11.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 238 |
+
blocks.11.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 239 |
+
blocks.11.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 240 |
+
blocks.11.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 241 |
+
blocks.11.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 242 |
+
blocks.11.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 243 |
+
blocks.12.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 244 |
+
blocks.12.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 245 |
+
blocks.12.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 246 |
+
blocks.12.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 247 |
+
blocks.12.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 248 |
+
blocks.12.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 249 |
+
blocks.12.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 250 |
+
blocks.12.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 251 |
+
blocks.12.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 252 |
+
blocks.12.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 253 |
+
blocks.12.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 254 |
+
blocks.12.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 255 |
+
blocks.12.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 256 |
+
blocks.12.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 257 |
+
blocks.12.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 258 |
+
blocks.12.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 259 |
+
blocks.12.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 260 |
+
blocks.12.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 261 |
+
blocks.13.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 262 |
+
blocks.13.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 263 |
+
blocks.13.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 264 |
+
blocks.13.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 265 |
+
blocks.13.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 266 |
+
blocks.13.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 267 |
+
blocks.13.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 268 |
+
blocks.13.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 269 |
+
blocks.13.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 270 |
+
blocks.13.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 271 |
+
blocks.13.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 272 |
+
blocks.13.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 273 |
+
blocks.13.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 274 |
+
blocks.13.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 275 |
+
blocks.13.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 276 |
+
blocks.13.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 277 |
+
blocks.13.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 278 |
+
blocks.13.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 279 |
+
blocks.14.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 280 |
+
blocks.14.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 281 |
+
blocks.14.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 282 |
+
blocks.14.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 283 |
+
blocks.14.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 284 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 285 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 286 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 287 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 288 |
+
blocks.14.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 289 |
+
blocks.14.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 290 |
+
blocks.14.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 291 |
+
blocks.14.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 292 |
+
blocks.14.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 293 |
+
blocks.14.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 294 |
+
blocks.14.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 295 |
+
blocks.14.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 296 |
+
blocks.14.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 297 |
+
blocks.14.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 298 |
+
blocks.14.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 299 |
+
blocks.14.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 300 |
+
blocks.14.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 301 |
+
blocks.15.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 302 |
+
blocks.15.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 303 |
+
blocks.15.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 304 |
+
blocks.15.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 305 |
+
blocks.15.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 306 |
+
blocks.15.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 307 |
+
blocks.15.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 308 |
+
blocks.15.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 309 |
+
blocks.15.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 310 |
+
blocks.15.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 311 |
+
blocks.15.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 312 |
+
blocks.15.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 313 |
+
blocks.15.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 314 |
+
blocks.15.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 315 |
+
blocks.15.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 316 |
+
blocks.15.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 317 |
+
blocks.15.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 318 |
+
blocks.15.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 319 |
+
blocks.16.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 320 |
+
blocks.16.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 321 |
+
blocks.16.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 322 |
+
blocks.16.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 323 |
+
blocks.16.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 324 |
+
blocks.16.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 325 |
+
blocks.16.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 326 |
+
blocks.16.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 327 |
+
blocks.16.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 328 |
+
blocks.16.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 329 |
+
blocks.16.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 330 |
+
blocks.16.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 331 |
+
blocks.16.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 332 |
+
blocks.16.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 333 |
+
blocks.16.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 334 |
+
blocks.16.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 335 |
+
blocks.16.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 336 |
+
blocks.16.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 337 |
+
blocks.17.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 338 |
+
blocks.17.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 339 |
+
blocks.17.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 340 |
+
blocks.17.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 341 |
+
blocks.17.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 342 |
+
blocks.17.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 343 |
+
blocks.17.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 344 |
+
blocks.17.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 345 |
+
blocks.17.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 346 |
+
blocks.17.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 347 |
+
blocks.17.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 348 |
+
blocks.17.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 349 |
+
blocks.17.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 350 |
+
blocks.17.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 351 |
+
blocks.17.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 352 |
+
blocks.17.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 353 |
+
blocks.17.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 354 |
+
blocks.17.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 355 |
+
norm_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 356 |
+
policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 357 |
+
policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 358 |
+
policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 359 |
+
policy_head.linear_g.weight, [48, 144], 6912 params
|
| 360 |
+
policy_head.linear_pass.weight, [6, 144], 864 params
|
| 361 |
+
policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 362 |
+
policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 363 |
+
value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 364 |
+
value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 365 |
+
value_head.linear2.weight, [128, 288], 36864 params
|
| 366 |
+
value_head.linear2.bias, [128], 128 params
|
| 367 |
+
value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 368 |
+
value_head.linear_valuehead.bias, [3], 3 params
|
| 369 |
+
value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 370 |
+
value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 371 |
+
value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 372 |
+
value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 373 |
+
value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 374 |
+
value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 375 |
+
value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 376 |
+
value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 377 |
+
value_head.linear_s2.weight, [112, 288], 32256 params
|
| 378 |
+
value_head.linear_s2.bias, [112], 112 params
|
| 379 |
+
value_head.linear_s2off.weight, [112, 1], 112 params
|
| 380 |
+
value_head.linear_s2par.weight, [112, 1], 112 params
|
| 381 |
+
value_head.linear_s3.weight, [8, 112], 896 params
|
| 382 |
+
value_head.linear_s3.bias, [8], 8 params
|
| 383 |
+
value_head.linear_smix.weight, [8, 288], 2304 params
|
| 384 |
+
value_head.linear_smix.bias, [8], 8 params
|
| 385 |
+
norm_intermediate_trunkfinal.gamma, [1, 384, 1, 1], 384 params
|
| 386 |
+
norm_intermediate_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 387 |
+
intermediate_policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 388 |
+
intermediate_policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 389 |
+
intermediate_policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 390 |
+
intermediate_policy_head.linear_g.weight, [48, 144], 6912 params
|
| 391 |
+
intermediate_policy_head.linear_pass.weight, [6, 144], 864 params
|
| 392 |
+
intermediate_policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 393 |
+
intermediate_policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 394 |
+
intermediate_value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 395 |
+
intermediate_value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 396 |
+
intermediate_value_head.linear2.weight, [128, 288], 36864 params
|
| 397 |
+
intermediate_value_head.linear2.bias, [128], 128 params
|
| 398 |
+
intermediate_value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 399 |
+
intermediate_value_head.linear_valuehead.bias, [3], 3 params
|
| 400 |
+
intermediate_value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 401 |
+
intermediate_value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 402 |
+
intermediate_value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 403 |
+
intermediate_value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 404 |
+
intermediate_value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 405 |
+
intermediate_value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 406 |
+
intermediate_value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 407 |
+
intermediate_value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 408 |
+
intermediate_value_head.linear_s2.weight, [112, 288], 32256 params
|
| 409 |
+
intermediate_value_head.linear_s2.bias, [112], 112 params
|
| 410 |
+
intermediate_value_head.linear_s2off.weight, [112, 1], 112 params
|
| 411 |
+
intermediate_value_head.linear_s2par.weight, [112, 1], 112 params
|
| 412 |
+
intermediate_value_head.linear_s3.weight, [8, 112], 896 params
|
| 413 |
+
intermediate_value_head.linear_s3.bias, [8], 8 params
|
| 414 |
+
intermediate_value_head.linear_smix.weight, [8, 288], 2304 params
|
| 415 |
+
intermediate_value_head.linear_smix.bias, [8], 8 params
|
| 416 |
+
Total num params: 26934250
|
| 417 |
+
Total trainable params: 26934250
|
| 418 |
+
Writing model
|
| 419 |
+
Exported at:
|
| 420 |
+
2026-06-02 05:54:06.610895 UTC
|
shusaku_export/metadata.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"global_step_samples": 592160, "total_num_data_rows": 94964, "extra_stats": {"sums": {"p0loss_sum": 24886.099369000174, "p1loss_sum": 0.0, "p0softloss_sum": 179281.54515505303, "p1softloss_sum": 0.0, "vloss_sum": 25307.206279465754, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 3.241025743105681, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 23144.072598817856, "pacc1_sum": 24128.3321521658, "vsquare_sum": 10894.419248002061, "wsum": 592160.0, "nsamp": 592160.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 175697.75230622827, "sekiweightscale_sum": 255999.997668692, "norm_input_batch": 162.95434825997387, "norm_input_noreg_batch": 8.439416288520711, "norm_normal_batch": 710.6205826198674, "norm_normal_gamma_batch": 525.0578548481758, "norm_noreg_batch": 275.9002427703873, "norm_output_batch": 250.47710202426407, "norm_output_noreg_batch": 57.16589113347058, "Ip0loss_sum": 30209.05298405604, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 184085.41998421642, "Ip1softloss_sum": 0.0, "Ivloss_sum": 23221.689188796478, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 1.83967369922428, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 23647.44314373596, "Ivsquare_sum": 9344.917107723008, "pslr_batch": 0.00031001001001001, "wdnormal_batch": 0.00047538691627703987, "gnorm_cap_batch": 7750.25025025025, "window_start_batch": 0.0, "window_end_batch": 294397.9059059059, "gnorm_batch": 933.4238007742733, "exgnorm_sum": 0.0}, "weights": {"p0loss_sum": 31999.9997085865, "p1loss_sum": 31999.9997085865, "p0softloss_sum": 31999.9997085865, "p1softloss_sum": 31999.9997085865, "vloss_sum": 31999.9997085865, "tdvloss1_sum": 31999.9997085865, "tdvloss2_sum": 31999.9997085865, "tdvloss3_sum": 31999.9997085865, "tdsloss_sum": 31999.9997085865, "oloss_sum": 31999.9997085865, "sloss_sum": 31999.9997085865, "fploss_sum": 31999.9997085865, "skloss_sum": 31999.9997085865, "smloss_sum": 31999.9997085865, "sbcdfloss_sum": 31999.9997085865, "sbpdfloss_sum": 31999.9997085865, "sdregloss_sum": 31999.9997085865, "leadloss_sum": 31999.9997085865, "vtimeloss_sum": 31999.9997085865, "evstloss_sum": 31999.9997085865, "esstloss_sum": 31999.9997085865, "qwlloss_sum": 31999.9997085865, "qscloss_sum": 31999.9997085865, "loss_sum": 31999.9997085865, "pacc1_sum": 31999.9997085865, "vsquare_sum": 31999.9997085865, "wsum": 592160.0, "nsamp": 592160.0, "ptentr_sum": 31999.9997085865, "ptsoftentr_sum": 31999.9997085865, "sekiweightscale_sum": 31999.9997085865, "norm_input_batch": 3.1001001001001, "norm_input_noreg_batch": 3.1001001001001, "norm_normal_batch": 3.1001001001001, "norm_normal_gamma_batch": 3.1001001001001, "norm_noreg_batch": 3.1001001001001, "norm_output_batch": 3.1001001001001, "norm_output_noreg_batch": 3.1001001001001, "Ip0loss_sum": 31999.9997085865, "Ip1loss_sum": 31999.9997085865, "Ip0softloss_sum": 31999.9997085865, "Ip1softloss_sum": 31999.9997085865, "Ivloss_sum": 31999.9997085865, "Itdvloss1_sum": 31999.9997085865, "Itdvloss2_sum": 31999.9997085865, "Itdvloss3_sum": 31999.9997085865, "Itdsloss_sum": 31999.9997085865, "Ioloss_sum": 31999.9997085865, "Isloss_sum": 31999.9997085865, "Ifploss_sum": 31999.9997085865, "Iskloss_sum": 31999.9997085865, "Ismloss_sum": 31999.9997085865, "Isbcdfloss_sum": 31999.9997085865, "Isbpdfloss_sum": 31999.9997085865, "Isdregloss_sum": 31999.9997085865, "Ileadloss_sum": 31999.9997085865, "Ivtimeloss_sum": 31999.9997085865, "Ievstloss_sum": 31999.9997085865, "Iesstloss_sum": 31999.9997085865, "Iqwlloss_sum": 31999.9997085865, "Iqscloss_sum": 31999.9997085865, "Ipacc1_sum": 31999.9997085865, "Ivsquare_sum": 31999.9997085865, "pslr_batch": 3.1001001001001, "wdnormal_batch": 3.1001001001001, "gnorm_cap_batch": 3.1001001001001, "window_start_batch": 3.1001001001001, "window_end_batch": 3.1001001001001, "gnorm_batch": 3.1001001001001, "exgnorm_sum": 31995.308549272642}, "last_val_metrics": {"sums": {"p0loss_sum": 1238.3617725372314, "p1loss_sum": 0.0, "p0softloss_sum": 4518.516876220703, "p1softloss_sum": 0.0, "vloss_sum": 632.6955082416534, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 0.1092256527626887, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 1448.2003021240234, "pacc1_sum": 452.0, "vsquare_sum": 362.2229745388031, "wsum": 800.0, "nsamp": 800.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 4392.44384765625, "sekiweightscale_sum": 5600.0, "norm_input_batch": 1.3142590522766113, "norm_input_noreg_batch": 0.06804606318473816, "norm_normal_batch": 5.727265930175782, "norm_normal_gamma_batch": 4.233429718017578, "norm_noreg_batch": 2.224753761291504, "norm_output_batch": 2.019840049743652, "norm_output_noreg_batch": 0.46099209785461426, "Ip0loss_sum": 1163.1206130981445, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 4638.191925048828, "Ip1softloss_sum": 0.0, "Ivloss_sum": 576.0568370819092, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 0.06329357656068169, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 449.0, "Ivsquare_sum": 305.071961581707, "nsamp_train": 569664.0, "wsum_train": 569664.0}, "weights": {"p0loss_sum": 800.0, "p1loss_sum": 800.0, "p0softloss_sum": 800.0, "p1softloss_sum": 800.0, "vloss_sum": 800.0, "tdvloss1_sum": 800.0, "tdvloss2_sum": 800.0, "tdvloss3_sum": 800.0, "tdsloss_sum": 800.0, "oloss_sum": 800.0, "sloss_sum": 800.0, "fploss_sum": 800.0, "skloss_sum": 800.0, "smloss_sum": 800.0, "sbcdfloss_sum": 800.0, "sbpdfloss_sum": 800.0, "sdregloss_sum": 800.0, "leadloss_sum": 800.0, "vtimeloss_sum": 800.0, "evstloss_sum": 800.0, "esstloss_sum": 800.0, "qwlloss_sum": 800.0, "qscloss_sum": 800.0, "loss_sum": 800.0, "pacc1_sum": 800.0, "vsquare_sum": 800.0, "wsum": 800.0, "nsamp": 800.0, "ptentr_sum": 800.0, "ptsoftentr_sum": 800.0, "sekiweightscale_sum": 800.0, "norm_input_batch": 0.025, "norm_input_noreg_batch": 0.025, "norm_normal_batch": 0.025, "norm_normal_gamma_batch": 0.025, "norm_noreg_batch": 0.025, "norm_output_batch": 0.025, "norm_output_noreg_batch": 0.025, "Ip0loss_sum": 800.0, "Ip1loss_sum": 800.0, "Ip0softloss_sum": 800.0, "Ip1softloss_sum": 800.0, "Ivloss_sum": 800.0, "Itdvloss1_sum": 800.0, "Itdvloss2_sum": 800.0, "Itdvloss3_sum": 800.0, "Itdsloss_sum": 800.0, "Ioloss_sum": 800.0, "Isloss_sum": 800.0, "Ifploss_sum": 800.0, "Iskloss_sum": 800.0, "Ismloss_sum": 800.0, "Isbcdfloss_sum": 800.0, "Isbpdfloss_sum": 800.0, "Isdregloss_sum": 800.0, "Ileadloss_sum": 800.0, "Ivtimeloss_sum": 800.0, "Ievstloss_sum": 800.0, "Iesstloss_sum": 800.0, "Iqwlloss_sum": 800.0, "Iqscloss_sum": 800.0, "Ipacc1_sum": 800.0, "Ivsquare_sum": 800.0, "nsamp_train": 569664.0, "wsum_train": 569664.0}}}}
|
shusaku_export/model.bin.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77228d08f31e7fc2d12877b56a90a4bd256daad66c96bb6a6472d3f097b293d0
|
| 3 |
+
size 99065801
|
takemiya_export/best_checkpoint.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d888e45d505fe87437f51e79b98e10d0b09cfa566ee275f74dfe0079970b517e
|
| 3 |
+
size 214678035
|
takemiya_export/log.txt
ADDED
|
@@ -0,0 +1,420 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
['/home/ran/go-arm-rl/KataGo.OLD/python/export_model_pytorch.py', '-checkpoint', '/home/ran/go-arm-rl/models/takemiya_75/best_checkpoint.ckpt', '-export-dir', '/home/ran/go-arm-rl/models/takemiya_75', '-model-name', 'takemiya_75', '-filename-prefix', 'model', '-export-14-as-15']
|
| 2 |
+
{'version': 14, 'norm_kind': 'fixscaleonenorm', 'bnorm_epsilon': 0.0001, 'bnorm_running_avg_momentum': 0.001, 'initial_conv_1x1': False, 'trunk_num_channels': 384, 'mid_num_channels': 192, 'gpool_num_channels': 64, 'use_attention_pool': False, 'num_attention_pool_heads': 4, 'block_kind': [['rconv1', 'bottlenest2'], ['rconv2', 'bottlenest2'], ['rconv3', 'bottlenest2gpool'], ['rconv4', 'bottlenest2'], ['rconv5', 'bottlenest2'], ['rconv6', 'bottlenest2gpool'], ['rconv7', 'bottlenest2'], ['rconv8', 'bottlenest2'], ['rconv9', 'bottlenest2gpool'], ['rconv10', 'bottlenest2'], ['rconv11', 'bottlenest2'], ['rconv12', 'bottlenest2gpool'], ['rconv13', 'bottlenest2'], ['rconv14', 'bottlenest2'], ['rconv15', 'bottlenest2gpool'], ['rconv16', 'bottlenest2'], ['rconv17', 'bottlenest2'], ['rconv18', 'bottlenest2']], 'p1_num_channels': 48, 'g1_num_channels': 48, 'v1_num_channels': 96, 'sbv2_num_channels': 112, 'num_scorebeliefs': 8, 'v2_size': 128, 'bnorm_use_gamma': True, 'activation': 'mish', 'use_repvgg_init': True, 'use_repvgg_learning_rate': True, 'has_intermediate_head': True, 'intermediate_head_blocks': 18, 'trunk_normless': True, 'metadata_encoder': {'internal_num_channels': 384}}
|
| 3 |
+
Parameters in model:
|
| 4 |
+
conv_spatial.weight, [384, 22, 3, 3], 76032 params
|
| 5 |
+
linear_global.weight, [384, 19], 7296 params
|
| 6 |
+
metadata_encoder.linear1.weight, [384, 192], 73728 params
|
| 7 |
+
metadata_encoder.linear1.bias, [384], 384 params
|
| 8 |
+
metadata_encoder.linear2.weight, [384, 384], 147456 params
|
| 9 |
+
metadata_encoder.linear2.bias, [384], 384 params
|
| 10 |
+
metadata_encoder.linear_output_to_trunk.weight, [384, 384], 147456 params
|
| 11 |
+
blocks.0.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 12 |
+
blocks.0.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 13 |
+
blocks.0.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 14 |
+
blocks.0.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 15 |
+
blocks.0.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 16 |
+
blocks.0.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 17 |
+
blocks.0.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 18 |
+
blocks.0.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 19 |
+
blocks.0.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 20 |
+
blocks.0.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 21 |
+
blocks.0.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 22 |
+
blocks.0.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 23 |
+
blocks.0.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 24 |
+
blocks.0.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 25 |
+
blocks.0.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 26 |
+
blocks.0.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 27 |
+
blocks.0.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 28 |
+
blocks.0.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 29 |
+
blocks.1.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 30 |
+
blocks.1.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 31 |
+
blocks.1.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 32 |
+
blocks.1.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 33 |
+
blocks.1.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 34 |
+
blocks.1.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 35 |
+
blocks.1.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 36 |
+
blocks.1.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 37 |
+
blocks.1.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 38 |
+
blocks.1.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 39 |
+
blocks.1.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 40 |
+
blocks.1.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 41 |
+
blocks.1.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 42 |
+
blocks.1.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 43 |
+
blocks.1.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 44 |
+
blocks.1.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 45 |
+
blocks.1.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 46 |
+
blocks.1.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 47 |
+
blocks.2.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 48 |
+
blocks.2.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 49 |
+
blocks.2.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 50 |
+
blocks.2.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 51 |
+
blocks.2.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 52 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 53 |
+
blocks.2.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 54 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 55 |
+
blocks.2.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 56 |
+
blocks.2.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 57 |
+
blocks.2.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 58 |
+
blocks.2.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 59 |
+
blocks.2.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 60 |
+
blocks.2.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 61 |
+
blocks.2.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 62 |
+
blocks.2.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 63 |
+
blocks.2.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 64 |
+
blocks.2.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 65 |
+
blocks.2.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 66 |
+
blocks.2.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 67 |
+
blocks.2.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 68 |
+
blocks.2.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 69 |
+
blocks.3.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 70 |
+
blocks.3.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 71 |
+
blocks.3.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 72 |
+
blocks.3.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 73 |
+
blocks.3.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 74 |
+
blocks.3.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 75 |
+
blocks.3.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 76 |
+
blocks.3.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 77 |
+
blocks.3.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 78 |
+
blocks.3.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 79 |
+
blocks.3.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 80 |
+
blocks.3.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 81 |
+
blocks.3.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 82 |
+
blocks.3.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 83 |
+
blocks.3.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 84 |
+
blocks.3.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 85 |
+
blocks.3.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 86 |
+
blocks.3.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 87 |
+
blocks.4.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 88 |
+
blocks.4.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 89 |
+
blocks.4.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 90 |
+
blocks.4.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 91 |
+
blocks.4.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 92 |
+
blocks.4.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 93 |
+
blocks.4.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 94 |
+
blocks.4.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 95 |
+
blocks.4.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 96 |
+
blocks.4.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 97 |
+
blocks.4.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 98 |
+
blocks.4.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 99 |
+
blocks.4.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 100 |
+
blocks.4.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 101 |
+
blocks.4.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 102 |
+
blocks.4.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 103 |
+
blocks.4.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 104 |
+
blocks.4.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 105 |
+
blocks.5.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 106 |
+
blocks.5.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 107 |
+
blocks.5.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 108 |
+
blocks.5.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 109 |
+
blocks.5.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 110 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 111 |
+
blocks.5.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 112 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 113 |
+
blocks.5.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 114 |
+
blocks.5.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 115 |
+
blocks.5.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 116 |
+
blocks.5.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 117 |
+
blocks.5.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 118 |
+
blocks.5.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 119 |
+
blocks.5.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 120 |
+
blocks.5.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 121 |
+
blocks.5.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 122 |
+
blocks.5.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 123 |
+
blocks.5.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 124 |
+
blocks.5.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 125 |
+
blocks.5.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 126 |
+
blocks.5.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 127 |
+
blocks.6.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 128 |
+
blocks.6.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 129 |
+
blocks.6.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 130 |
+
blocks.6.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 131 |
+
blocks.6.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 132 |
+
blocks.6.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 133 |
+
blocks.6.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 134 |
+
blocks.6.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 135 |
+
blocks.6.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 136 |
+
blocks.6.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 137 |
+
blocks.6.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 138 |
+
blocks.6.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 139 |
+
blocks.6.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 140 |
+
blocks.6.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 141 |
+
blocks.6.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 142 |
+
blocks.6.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 143 |
+
blocks.6.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 144 |
+
blocks.6.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 145 |
+
blocks.7.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 146 |
+
blocks.7.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 147 |
+
blocks.7.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 148 |
+
blocks.7.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 149 |
+
blocks.7.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 150 |
+
blocks.7.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 151 |
+
blocks.7.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 152 |
+
blocks.7.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 153 |
+
blocks.7.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 154 |
+
blocks.7.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 155 |
+
blocks.7.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 156 |
+
blocks.7.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 157 |
+
blocks.7.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 158 |
+
blocks.7.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 159 |
+
blocks.7.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 160 |
+
blocks.7.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 161 |
+
blocks.7.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 162 |
+
blocks.7.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 163 |
+
blocks.8.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 164 |
+
blocks.8.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 165 |
+
blocks.8.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 166 |
+
blocks.8.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 167 |
+
blocks.8.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 168 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 169 |
+
blocks.8.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 170 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 171 |
+
blocks.8.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 172 |
+
blocks.8.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 173 |
+
blocks.8.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 174 |
+
blocks.8.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 175 |
+
blocks.8.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 176 |
+
blocks.8.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 177 |
+
blocks.8.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 178 |
+
blocks.8.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 179 |
+
blocks.8.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 180 |
+
blocks.8.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 181 |
+
blocks.8.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 182 |
+
blocks.8.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 183 |
+
blocks.8.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 184 |
+
blocks.8.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 185 |
+
blocks.9.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 186 |
+
blocks.9.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 187 |
+
blocks.9.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 188 |
+
blocks.9.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 189 |
+
blocks.9.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 190 |
+
blocks.9.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 191 |
+
blocks.9.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 192 |
+
blocks.9.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 193 |
+
blocks.9.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 194 |
+
blocks.9.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 195 |
+
blocks.9.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 196 |
+
blocks.9.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 197 |
+
blocks.9.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 198 |
+
blocks.9.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 199 |
+
blocks.9.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 200 |
+
blocks.9.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 201 |
+
blocks.9.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 202 |
+
blocks.9.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 203 |
+
blocks.10.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 204 |
+
blocks.10.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 205 |
+
blocks.10.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 206 |
+
blocks.10.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 207 |
+
blocks.10.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 208 |
+
blocks.10.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 209 |
+
blocks.10.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 210 |
+
blocks.10.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 211 |
+
blocks.10.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 212 |
+
blocks.10.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 213 |
+
blocks.10.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 214 |
+
blocks.10.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 215 |
+
blocks.10.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 216 |
+
blocks.10.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 217 |
+
blocks.10.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 218 |
+
blocks.10.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 219 |
+
blocks.10.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 220 |
+
blocks.10.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 221 |
+
blocks.11.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 222 |
+
blocks.11.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 223 |
+
blocks.11.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 224 |
+
blocks.11.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 225 |
+
blocks.11.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 226 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 227 |
+
blocks.11.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 228 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 229 |
+
blocks.11.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 230 |
+
blocks.11.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 231 |
+
blocks.11.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 232 |
+
blocks.11.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 233 |
+
blocks.11.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 234 |
+
blocks.11.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 235 |
+
blocks.11.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 236 |
+
blocks.11.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 237 |
+
blocks.11.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 238 |
+
blocks.11.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 239 |
+
blocks.11.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 240 |
+
blocks.11.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 241 |
+
blocks.11.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 242 |
+
blocks.11.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 243 |
+
blocks.12.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 244 |
+
blocks.12.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 245 |
+
blocks.12.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 246 |
+
blocks.12.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 247 |
+
blocks.12.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 248 |
+
blocks.12.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 249 |
+
blocks.12.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 250 |
+
blocks.12.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 251 |
+
blocks.12.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 252 |
+
blocks.12.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 253 |
+
blocks.12.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 254 |
+
blocks.12.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 255 |
+
blocks.12.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 256 |
+
blocks.12.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 257 |
+
blocks.12.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 258 |
+
blocks.12.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 259 |
+
blocks.12.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 260 |
+
blocks.12.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 261 |
+
blocks.13.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 262 |
+
blocks.13.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 263 |
+
blocks.13.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 264 |
+
blocks.13.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 265 |
+
blocks.13.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 266 |
+
blocks.13.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 267 |
+
blocks.13.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 268 |
+
blocks.13.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 269 |
+
blocks.13.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 270 |
+
blocks.13.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 271 |
+
blocks.13.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 272 |
+
blocks.13.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 273 |
+
blocks.13.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 274 |
+
blocks.13.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 275 |
+
blocks.13.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 276 |
+
blocks.13.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 277 |
+
blocks.13.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 278 |
+
blocks.13.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 279 |
+
blocks.14.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 280 |
+
blocks.14.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 281 |
+
blocks.14.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 282 |
+
blocks.14.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 283 |
+
blocks.14.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 284 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1r.weight, [128, 192, 3, 3], 221184 params
|
| 285 |
+
blocks.14.blockstack.0.normactconv1.convpool.conv1g.weight, [64, 192, 3, 3], 110592 params
|
| 286 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.beta, [1, 64, 1, 1], 64 params
|
| 287 |
+
blocks.14.blockstack.0.normactconv1.convpool.normg.gamma, [1, 64, 1, 1], 64 params
|
| 288 |
+
blocks.14.blockstack.0.normactconv1.convpool.linear_g.weight, [128, 192], 24576 params
|
| 289 |
+
blocks.14.blockstack.0.normactconv2.norm.beta, [1, 128, 1, 1], 128 params
|
| 290 |
+
blocks.14.blockstack.0.normactconv2.norm.gamma, [1, 128, 1, 1], 128 params
|
| 291 |
+
blocks.14.blockstack.0.normactconv2.conv.weight, [192, 128, 3, 3], 221184 params
|
| 292 |
+
blocks.14.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 293 |
+
blocks.14.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 294 |
+
blocks.14.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 295 |
+
blocks.14.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 296 |
+
blocks.14.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 297 |
+
blocks.14.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 298 |
+
blocks.14.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 299 |
+
blocks.14.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 300 |
+
blocks.14.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 301 |
+
blocks.15.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 302 |
+
blocks.15.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 303 |
+
blocks.15.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 304 |
+
blocks.15.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 305 |
+
blocks.15.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 306 |
+
blocks.15.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 307 |
+
blocks.15.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 308 |
+
blocks.15.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 309 |
+
blocks.15.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 310 |
+
blocks.15.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 311 |
+
blocks.15.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 312 |
+
blocks.15.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 313 |
+
blocks.15.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 314 |
+
blocks.15.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 315 |
+
blocks.15.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 316 |
+
blocks.15.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 317 |
+
blocks.15.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 318 |
+
blocks.15.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 319 |
+
blocks.16.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 320 |
+
blocks.16.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 321 |
+
blocks.16.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 322 |
+
blocks.16.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 323 |
+
blocks.16.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 324 |
+
blocks.16.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 325 |
+
blocks.16.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 326 |
+
blocks.16.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 327 |
+
blocks.16.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 328 |
+
blocks.16.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 329 |
+
blocks.16.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 330 |
+
blocks.16.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 331 |
+
blocks.16.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 332 |
+
blocks.16.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 333 |
+
blocks.16.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 334 |
+
blocks.16.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 335 |
+
blocks.16.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 336 |
+
blocks.16.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 337 |
+
blocks.17.normactconvp.norm.beta, [1, 384, 1, 1], 384 params
|
| 338 |
+
blocks.17.normactconvp.norm.gamma, [1, 384, 1, 1], 384 params
|
| 339 |
+
blocks.17.normactconvp.conv.weight, [192, 384, 1, 1], 73728 params
|
| 340 |
+
blocks.17.blockstack.0.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 341 |
+
blocks.17.blockstack.0.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 342 |
+
blocks.17.blockstack.0.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 343 |
+
blocks.17.blockstack.0.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 344 |
+
blocks.17.blockstack.0.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 345 |
+
blocks.17.blockstack.0.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 346 |
+
blocks.17.blockstack.1.normactconv1.norm.beta, [1, 192, 1, 1], 192 params
|
| 347 |
+
blocks.17.blockstack.1.normactconv1.norm.gamma, [1, 192, 1, 1], 192 params
|
| 348 |
+
blocks.17.blockstack.1.normactconv1.conv.weight, [192, 192, 3, 3], 331776 params
|
| 349 |
+
blocks.17.blockstack.1.normactconv2.norm.beta, [1, 192, 1, 1], 192 params
|
| 350 |
+
blocks.17.blockstack.1.normactconv2.norm.gamma, [1, 192, 1, 1], 192 params
|
| 351 |
+
blocks.17.blockstack.1.normactconv2.conv.weight, [192, 192, 3, 3], 331776 params
|
| 352 |
+
blocks.17.normactconvq.norm.beta, [1, 192, 1, 1], 192 params
|
| 353 |
+
blocks.17.normactconvq.norm.gamma, [1, 192, 1, 1], 192 params
|
| 354 |
+
blocks.17.normactconvq.conv.weight, [384, 192, 1, 1], 73728 params
|
| 355 |
+
norm_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 356 |
+
policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 357 |
+
policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 358 |
+
policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 359 |
+
policy_head.linear_g.weight, [48, 144], 6912 params
|
| 360 |
+
policy_head.linear_pass.weight, [6, 144], 864 params
|
| 361 |
+
policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 362 |
+
policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 363 |
+
value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 364 |
+
value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 365 |
+
value_head.linear2.weight, [128, 288], 36864 params
|
| 366 |
+
value_head.linear2.bias, [128], 128 params
|
| 367 |
+
value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 368 |
+
value_head.linear_valuehead.bias, [3], 3 params
|
| 369 |
+
value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 370 |
+
value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 371 |
+
value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 372 |
+
value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 373 |
+
value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 374 |
+
value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 375 |
+
value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 376 |
+
value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 377 |
+
value_head.linear_s2.weight, [112, 288], 32256 params
|
| 378 |
+
value_head.linear_s2.bias, [112], 112 params
|
| 379 |
+
value_head.linear_s2off.weight, [112, 1], 112 params
|
| 380 |
+
value_head.linear_s2par.weight, [112, 1], 112 params
|
| 381 |
+
value_head.linear_s3.weight, [8, 112], 896 params
|
| 382 |
+
value_head.linear_s3.bias, [8], 8 params
|
| 383 |
+
value_head.linear_smix.weight, [8, 288], 2304 params
|
| 384 |
+
value_head.linear_smix.bias, [8], 8 params
|
| 385 |
+
norm_intermediate_trunkfinal.gamma, [1, 384, 1, 1], 384 params
|
| 386 |
+
norm_intermediate_trunkfinal.beta, [1, 384, 1, 1], 384 params
|
| 387 |
+
intermediate_policy_head.conv1p.weight, [48, 384, 1, 1], 18432 params
|
| 388 |
+
intermediate_policy_head.conv1g.weight, [48, 384, 1, 1], 18432 params
|
| 389 |
+
intermediate_policy_head.biasg.beta, [1, 48, 1, 1], 48 params
|
| 390 |
+
intermediate_policy_head.linear_g.weight, [48, 144], 6912 params
|
| 391 |
+
intermediate_policy_head.linear_pass.weight, [6, 144], 864 params
|
| 392 |
+
intermediate_policy_head.bias2.beta, [1, 48, 1, 1], 48 params
|
| 393 |
+
intermediate_policy_head.conv2p.weight, [6, 48, 1, 1], 288 params
|
| 394 |
+
intermediate_value_head.conv1.weight, [96, 384, 1, 1], 36864 params
|
| 395 |
+
intermediate_value_head.bias1.beta, [1, 96, 1, 1], 96 params
|
| 396 |
+
intermediate_value_head.linear2.weight, [128, 288], 36864 params
|
| 397 |
+
intermediate_value_head.linear2.bias, [128], 128 params
|
| 398 |
+
intermediate_value_head.linear_valuehead.weight, [3, 128], 384 params
|
| 399 |
+
intermediate_value_head.linear_valuehead.bias, [3], 3 params
|
| 400 |
+
intermediate_value_head.linear_miscvaluehead.weight, [10, 128], 1280 params
|
| 401 |
+
intermediate_value_head.linear_miscvaluehead.bias, [10], 10 params
|
| 402 |
+
intermediate_value_head.linear_moremiscvaluehead.weight, [8, 128], 1024 params
|
| 403 |
+
intermediate_value_head.linear_moremiscvaluehead.bias, [8], 8 params
|
| 404 |
+
intermediate_value_head.conv_ownership.weight, [1, 96, 1, 1], 96 params
|
| 405 |
+
intermediate_value_head.conv_scoring.weight, [1, 96, 1, 1], 96 params
|
| 406 |
+
intermediate_value_head.conv_futurepos.weight, [2, 384, 1, 1], 768 params
|
| 407 |
+
intermediate_value_head.conv_seki.weight, [4, 384, 1, 1], 1536 params
|
| 408 |
+
intermediate_value_head.linear_s2.weight, [112, 288], 32256 params
|
| 409 |
+
intermediate_value_head.linear_s2.bias, [112], 112 params
|
| 410 |
+
intermediate_value_head.linear_s2off.weight, [112, 1], 112 params
|
| 411 |
+
intermediate_value_head.linear_s2par.weight, [112, 1], 112 params
|
| 412 |
+
intermediate_value_head.linear_s3.weight, [8, 112], 896 params
|
| 413 |
+
intermediate_value_head.linear_s3.bias, [8], 8 params
|
| 414 |
+
intermediate_value_head.linear_smix.weight, [8, 288], 2304 params
|
| 415 |
+
intermediate_value_head.linear_smix.bias, [8], 8 params
|
| 416 |
+
Total num params: 26934250
|
| 417 |
+
Total trainable params: 26934250
|
| 418 |
+
Writing model
|
| 419 |
+
Exported at:
|
| 420 |
+
2026-06-02 05:54:17.925720 UTC
|
takemiya_export/metadata.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"global_step_samples": 5655840, "total_num_data_rows": 377064, "extra_stats": {"sums": {"p0loss_sum": 26079.318906033994, "p1loss_sum": 0.0, "p0softloss_sum": 183745.02191034125, "p1softloss_sum": 0.0, "vloss_sum": 22697.494352537087, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 2.930158174198707, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 24253.766769468217, "pacc1_sum": 23445.055070156563, "vsquare_sum": 13868.767374340494, "wsum": 5655840.0, "nsamp": 5655840.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 175697.7539062353, "sekiweightscale_sum": 255999.99999998524, "norm_input_batch": 3307.4913539305758, "norm_input_noreg_batch": 176.42830619509772, "norm_normal_batch": 15432.564331413778, "norm_normal_gamma_batch": 10897.867222456842, "norm_noreg_batch": 5761.776472803008, "norm_output_batch": 5185.646029537749, "norm_output_noreg_batch": 1181.6277524103293, "Ip0loss_sum": 31489.51554698905, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 183672.11185634372, "Ip1softloss_sum": 0.0, "Ivloss_sum": 19934.36035324671, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 1.919196658692693, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 22945.881731047375, "Ivsquare_sum": 9361.359613187587, "pslr_batch": 0.006410010010010014, "wdnormal_batch": 0.01088345986366978, "gnorm_cap_batch": 160250.25025025025, "window_start_batch": 0.0, "window_end_batch": 24169840.144144144, "gnorm_batch": 18810.850871452603, "exgnorm_sum": 0.0}, "weights": {"p0loss_sum": 31999.999999998156, "p1loss_sum": 31999.999999998156, "p0softloss_sum": 31999.999999998156, "p1softloss_sum": 31999.999999998156, "vloss_sum": 31999.999999998156, "tdvloss1_sum": 31999.999999998156, "tdvloss2_sum": 31999.999999998156, "tdvloss3_sum": 31999.999999998156, "tdsloss_sum": 31999.999999998156, "oloss_sum": 31999.999999998156, "sloss_sum": 31999.999999998156, "fploss_sum": 31999.999999998156, "skloss_sum": 31999.999999998156, "smloss_sum": 31999.999999998156, "sbcdfloss_sum": 31999.999999998156, "sbpdfloss_sum": 31999.999999998156, "sdregloss_sum": 31999.999999998156, "leadloss_sum": 31999.999999998156, "vtimeloss_sum": 31999.999999998156, "evstloss_sum": 31999.999999998156, "esstloss_sum": 31999.999999998156, "qwlloss_sum": 31999.999999998156, "qscloss_sum": 31999.999999998156, "loss_sum": 31999.999999998156, "pacc1_sum": 31999.999999998156, "vsquare_sum": 31999.999999998156, "wsum": 5655840.0, "nsamp": 5655840.0, "ptentr_sum": 31999.999999998156, "ptsoftentr_sum": 31999.999999998156, "sekiweightscale_sum": 31999.999999998156, "norm_input_batch": 64.1001001001001, "norm_input_noreg_batch": 64.1001001001001, "norm_normal_batch": 64.1001001001001, "norm_normal_gamma_batch": 64.1001001001001, "norm_noreg_batch": 64.1001001001001, "norm_output_batch": 64.1001001001001, "norm_output_noreg_batch": 64.1001001001001, "Ip0loss_sum": 31999.999999998156, "Ip1loss_sum": 31999.999999998156, "Ip0softloss_sum": 31999.999999998156, "Ip1softloss_sum": 31999.999999998156, "Ivloss_sum": 31999.999999998156, "Itdvloss1_sum": 31999.999999998156, "Itdvloss2_sum": 31999.999999998156, "Itdvloss3_sum": 31999.999999998156, "Itdsloss_sum": 31999.999999998156, "Ioloss_sum": 31999.999999998156, "Isloss_sum": 31999.999999998156, "Ifploss_sum": 31999.999999998156, "Iskloss_sum": 31999.999999998156, "Ismloss_sum": 31999.999999998156, "Isbcdfloss_sum": 31999.999999998156, "Isbpdfloss_sum": 31999.999999998156, "Isdregloss_sum": 31999.999999998156, "Ileadloss_sum": 31999.999999998156, "Ivtimeloss_sum": 31999.999999998156, "Ievstloss_sum": 31999.999999998156, "Iesstloss_sum": 31999.999999998156, "Iqwlloss_sum": 31999.999999998156, "Iqscloss_sum": 31999.999999998156, "Ipacc1_sum": 31999.999999998156, "Ivsquare_sum": 31999.999999998156, "pslr_batch": 64.1001001001001, "wdnormal_batch": 64.1001001001001, "gnorm_cap_batch": 64.1001001001001, "window_start_batch": 64.1001001001001, "window_end_batch": 64.1001001001001, "gnorm_batch": 64.1001001001001, "exgnorm_sum": 31995.188160531787}, "last_val_metrics": {"sums": {"p0loss_sum": 1745.6093826293945, "p1loss_sum": 0.0, "p0softloss_sum": 5397.02473449707, "p1softloss_sum": 0.0, "vloss_sum": 542.8788409233093, "tdvloss1_sum": 0.0, "tdvloss2_sum": 0.0, "tdvloss3_sum": 0.0, "tdsloss_sum": 0.0, "oloss_sum": 0.0, "sloss_sum": 0.0, "fploss_sum": 0.0, "skloss_sum": 0.0, "smloss_sum": 0.0, "sbcdfloss_sum": 0.0, "sbpdfloss_sum": 0.0, "sdregloss_sum": 0.09807993320282549, "leadloss_sum": 0.0, "vtimeloss_sum": 0.0, "evstloss_sum": 0.0, "esstloss_sum": 0.0, "qwlloss_sum": 0.0, "qscloss_sum": 0.0, "loss_sum": 1713.603572845459, "pacc1_sum": 495.0, "vsquare_sum": 355.3854830265045, "wsum": 928.0, "nsamp": 928.0, "ptentr_sum": 0.0, "ptsoftentr_sum": 5095.23486328125, "sekiweightscale_sum": 6496.0, "norm_input_batch": 1.4967471008300781, "norm_input_noreg_batch": 0.07980635023117065, "norm_normal_batch": 6.9749964904785156, "norm_normal_gamma_batch": 4.93140495300293, "norm_noreg_batch": 2.6061128158569336, "norm_output_batch": 2.346004463195801, "norm_output_noreg_batch": 0.5345845069885254, "Ip0loss_sum": 1445.8392715454102, "Ip1loss_sum": 0.0, "Ip0softloss_sum": 5354.578903198242, "Ip1softloss_sum": 0.0, "Ivloss_sum": 516.6796987056732, "Itdvloss1_sum": 0.0, "Itdvloss2_sum": 0.0, "Itdvloss3_sum": 0.0, "Itdsloss_sum": 0.0, "Ioloss_sum": 0.0, "Isloss_sum": 0.0, "Ifploss_sum": 0.0, "Iskloss_sum": 0.0, "Ismloss_sum": 0.0, "Isbcdfloss_sum": 0.0, "Isbpdfloss_sum": 0.0, "Isdregloss_sum": 0.05550521190161817, "Ileadloss_sum": 0.0, "Ivtimeloss_sum": 0.0, "Ievstloss_sum": 0.0, "Iesstloss_sum": 0.0, "Iqwlloss_sum": 0.0, "Iqscloss_sum": 0.0, "Ipacc1_sum": 502.0, "Ivsquare_sum": 261.4623956680298, "nsamp_train": 5583392.0, "wsum_train": 5583392.0}, "weights": {"p0loss_sum": 928.0, "p1loss_sum": 928.0, "p0softloss_sum": 928.0, "p1softloss_sum": 928.0, "vloss_sum": 928.0, "tdvloss1_sum": 928.0, "tdvloss2_sum": 928.0, "tdvloss3_sum": 928.0, "tdsloss_sum": 928.0, "oloss_sum": 928.0, "sloss_sum": 928.0, "fploss_sum": 928.0, "skloss_sum": 928.0, "smloss_sum": 928.0, "sbcdfloss_sum": 928.0, "sbpdfloss_sum": 928.0, "sdregloss_sum": 928.0, "leadloss_sum": 928.0, "vtimeloss_sum": 928.0, "evstloss_sum": 928.0, "esstloss_sum": 928.0, "qwlloss_sum": 928.0, "qscloss_sum": 928.0, "loss_sum": 928.0, "pacc1_sum": 928.0, "vsquare_sum": 928.0, "wsum": 928.0, "nsamp": 928.0, "ptentr_sum": 928.0, "ptsoftentr_sum": 928.0, "sekiweightscale_sum": 928.0, "norm_input_batch": 0.029, "norm_input_noreg_batch": 0.029, "norm_normal_batch": 0.029, "norm_normal_gamma_batch": 0.029, "norm_noreg_batch": 0.029, "norm_output_batch": 0.029, "norm_output_noreg_batch": 0.029, "Ip0loss_sum": 928.0, "Ip1loss_sum": 928.0, "Ip0softloss_sum": 928.0, "Ip1softloss_sum": 928.0, "Ivloss_sum": 928.0, "Itdvloss1_sum": 928.0, "Itdvloss2_sum": 928.0, "Itdvloss3_sum": 928.0, "Itdsloss_sum": 928.0, "Ioloss_sum": 928.0, "Isloss_sum": 928.0, "Ifploss_sum": 928.0, "Iskloss_sum": 928.0, "Ismloss_sum": 928.0, "Isbcdfloss_sum": 928.0, "Isbpdfloss_sum": 928.0, "Isdregloss_sum": 928.0, "Ileadloss_sum": 928.0, "Ivtimeloss_sum": 928.0, "Ievstloss_sum": 928.0, "Iesstloss_sum": 928.0, "Iqwlloss_sum": 928.0, "Iqscloss_sum": 928.0, "Ipacc1_sum": 928.0, "Ivsquare_sum": 928.0, "nsamp_train": 5583392.0, "wsum_train": 5583392.0}}}}
|
takemiya_export/model.bin.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ee8ca63701062f6ae22675027d197a2ec7f5559962aaeba2cdef52142dd1475
|
| 3 |
+
size 99059064
|