Upload 2 files
#2
by suad0 - opened
- config.json +84 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"num_blocks": 12,
|
| 3 |
+
"embedding_dim": 512,
|
| 4 |
+
"input_patch_size": 32,
|
| 5 |
+
"output_patch_size": 32,
|
| 6 |
+
"input_ff_dim": 2048,
|
| 7 |
+
"h_expand": 4,
|
| 8 |
+
"context_len": 2048,
|
| 9 |
+
"future_len": 320,
|
| 10 |
+
"dropout": 0.1,
|
| 11 |
+
"quantiles": [
|
| 12 |
+
0.1,
|
| 13 |
+
0.2,
|
| 14 |
+
0.3,
|
| 15 |
+
0.4,
|
| 16 |
+
0.5,
|
| 17 |
+
0.6,
|
| 18 |
+
0.7,
|
| 19 |
+
0.8,
|
| 20 |
+
0.9
|
| 21 |
+
],
|
| 22 |
+
"act_func": "torch.nn.SiLU",
|
| 23 |
+
"use_qk_norm": true,
|
| 24 |
+
"tokenizer_cfg": {
|
| 25 |
+
"input_patch_size": 32,
|
| 26 |
+
"output_patch_size": 32,
|
| 27 |
+
"input_patch_stride": 32
|
| 28 |
+
},
|
| 29 |
+
"scaler_cfg": {
|
| 30 |
+
"use_arcsinh": true,
|
| 31 |
+
"binaryaware": true
|
| 32 |
+
},
|
| 33 |
+
"stack_out_norm_config": {
|
| 34 |
+
"eps": 1e-07
|
| 35 |
+
},
|
| 36 |
+
"tta_diff": true,
|
| 37 |
+
"stack_config": {
|
| 38 |
+
"templates": {
|
| 39 |
+
"slstm_attn": {
|
| 40 |
+
"time_mixer": {
|
| 41 |
+
"model_type": "bi-slstm",
|
| 42 |
+
"embedding_dim": 512,
|
| 43 |
+
"num_heads": 4,
|
| 44 |
+
"rnn_backend": "flashrnn-cuda",
|
| 45 |
+
"num_slstm_heads": 4
|
| 46 |
+
},
|
| 47 |
+
"variate_mixer": {
|
| 48 |
+
"embedding_dim": 512,
|
| 49 |
+
"num_heads": 4
|
| 50 |
+
},
|
| 51 |
+
"dropout": 0.1,
|
| 52 |
+
"eps": 1e-07
|
| 53 |
+
},
|
| 54 |
+
"mlstm_attn": {
|
| 55 |
+
"time_mixer": {
|
| 56 |
+
"model_type": "bi-mlstm",
|
| 57 |
+
"embedding_dim": 512,
|
| 58 |
+
"num_heads": 4,
|
| 59 |
+
"rnn_backend": "flashrnn-cuda"
|
| 60 |
+
},
|
| 61 |
+
"variate_mixer": {
|
| 62 |
+
"embedding_dim": 512,
|
| 63 |
+
"num_heads": 4
|
| 64 |
+
},
|
| 65 |
+
"dropout": 0.1,
|
| 66 |
+
"eps": 1e-07
|
| 67 |
+
}
|
| 68 |
+
},
|
| 69 |
+
"recipe": [
|
| 70 |
+
"mlstm_attn",
|
| 71 |
+
"slstm_attn",
|
| 72 |
+
"mlstm_attn",
|
| 73 |
+
"slstm_attn",
|
| 74 |
+
"mlstm_attn",
|
| 75 |
+
"slstm_attn",
|
| 76 |
+
"mlstm_attn",
|
| 77 |
+
"slstm_attn",
|
| 78 |
+
"mlstm_attn",
|
| 79 |
+
"slstm_attn",
|
| 80 |
+
"mlstm_attn",
|
| 81 |
+
"slstm_attn"
|
| 82 |
+
]
|
| 83 |
+
}
|
| 84 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17760e4eee4c14e0d8bad8b9d9ce6f5076aaeeb9d956a88b8e95cf1493a1c176
|
| 3 |
+
size 330038664
|