Upload folder using huggingface_hub
Browse files- 000000000250/config.json +111 -0
- 000000000250/model.pt +3 -0
000000000250/config.json
ADDED
|
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"type": "slither.SlitherModel",
|
| 3 |
+
"pretrained_url": null,
|
| 4 |
+
"pretrained_step": null,
|
| 5 |
+
"pretrained_strict": null,
|
| 6 |
+
"torch_dtype": "float32",
|
| 7 |
+
"vocab_size": 49153,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"pad_token_id": 49152,
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"num_hidden_layers": 16,
|
| 13 |
+
"num_attention_heads": 8,
|
| 14 |
+
"num_key_value_heads": 4,
|
| 15 |
+
"intermediate_size": 1664,
|
| 16 |
+
"hidden_act": "silu",
|
| 17 |
+
"max_position_embeddings": 4096,
|
| 18 |
+
"rope_theta": 10000,
|
| 19 |
+
"rms_norm_eps": 1e-05,
|
| 20 |
+
"attention_kernel": "flash_attention",
|
| 21 |
+
"pure_modules": [
|
| 22 |
+
"SlitherStateWriter"
|
| 23 |
+
],
|
| 24 |
+
"position_embedding_dim": 256,
|
| 25 |
+
"num_output_layers": 4,
|
| 26 |
+
"num_memory_layers": 4,
|
| 27 |
+
"state_size": 512,
|
| 28 |
+
"num_state_heads": 16,
|
| 29 |
+
"init_state_out_scale": 1.0,
|
| 30 |
+
"write_mixer_kernel_size": 5,
|
| 31 |
+
"write_mixer_init_scale": 0.1,
|
| 32 |
+
"chunk_length": 1024,
|
| 33 |
+
"sharding": {
|
| 34 |
+
"replicate_default": true,
|
| 35 |
+
"backbone_layers.layers.*": [
|
| 36 |
+
[
|
| 37 |
+
"data",
|
| 38 |
+
"fsdp"
|
| 39 |
+
],
|
| 40 |
+
null,
|
| 41 |
+
null
|
| 42 |
+
],
|
| 43 |
+
"output_layers.layers.*": [
|
| 44 |
+
[
|
| 45 |
+
"data",
|
| 46 |
+
"fsdp"
|
| 47 |
+
],
|
| 48 |
+
null,
|
| 49 |
+
null
|
| 50 |
+
],
|
| 51 |
+
"memory_layers.layers.*": [
|
| 52 |
+
[
|
| 53 |
+
"data",
|
| 54 |
+
"fsdp"
|
| 55 |
+
],
|
| 56 |
+
null,
|
| 57 |
+
null
|
| 58 |
+
],
|
| 59 |
+
"lm_head": [
|
| 60 |
+
[
|
| 61 |
+
"data",
|
| 62 |
+
"fsdp"
|
| 63 |
+
],
|
| 64 |
+
null,
|
| 65 |
+
null
|
| 66 |
+
]
|
| 67 |
+
},
|
| 68 |
+
"remat": {
|
| 69 |
+
"advanced": [
|
| 70 |
+
{
|
| 71 |
+
"name": "backbone_layers",
|
| 72 |
+
"settings": {
|
| 73 |
+
"activation_checkpoint_layers": [
|
| 74 |
+
"SlitherBackboneLayer"
|
| 75 |
+
],
|
| 76 |
+
"optimization_barrier_layers": [
|
| 77 |
+
"SlitherBackboneLayer"
|
| 78 |
+
],
|
| 79 |
+
"scan_layers": "layers",
|
| 80 |
+
"is_layer_pure": true
|
| 81 |
+
}
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"name": "output_layers",
|
| 85 |
+
"settings": {
|
| 86 |
+
"activation_checkpoint_layers": [
|
| 87 |
+
"SlitherOutputLayer"
|
| 88 |
+
],
|
| 89 |
+
"optimization_barrier_layers": [
|
| 90 |
+
"SlitherOutputLayer"
|
| 91 |
+
],
|
| 92 |
+
"scan_layers": "layers",
|
| 93 |
+
"is_layer_pure": true
|
| 94 |
+
}
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"name": "memory_layers",
|
| 98 |
+
"settings": {
|
| 99 |
+
"activation_checkpoint_layers": [
|
| 100 |
+
"SlitherMemoryLayer"
|
| 101 |
+
],
|
| 102 |
+
"optimization_barrier_layers": [
|
| 103 |
+
"SlitherMemoryLayer"
|
| 104 |
+
],
|
| 105 |
+
"scan_layers": "layers",
|
| 106 |
+
"is_layer_pure": true
|
| 107 |
+
}
|
| 108 |
+
}
|
| 109 |
+
]
|
| 110 |
+
}
|
| 111 |
+
}
|
000000000250/model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d2bc9c066e7c5c08de3e162624c4b74fb7014832af718cbd2ecc74225026601
|
| 3 |
+
size 1529268251
|