Push model using huggingface_hub.
Browse files- config.json +1 -8
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -1,8 +1,5 @@
|
|
| 1 |
{
|
| 2 |
"activation": "gelu",
|
| 3 |
-
"architectures": [
|
| 4 |
-
"PretrainedDualAttnTransformerLM"
|
| 5 |
-
],
|
| 6 |
"bias": false,
|
| 7 |
"d_model": 1024,
|
| 8 |
"dff": null,
|
|
@@ -12,7 +9,6 @@
|
|
| 12 |
"n_heads_sa": 8,
|
| 13 |
"n_layers": 24,
|
| 14 |
"norm_first": true,
|
| 15 |
-
"norm_type": "layernorm",
|
| 16 |
"pos_enc_type": "RoPE",
|
| 17 |
"ra_kwargs": {
|
| 18 |
"n_kv_heads": 4,
|
|
@@ -25,7 +21,6 @@
|
|
| 25 |
"sa_kwargs": {
|
| 26 |
"n_kv_heads": 4
|
| 27 |
},
|
| 28 |
-
"share_attn_params": false,
|
| 29 |
"symbol_retrieval": "symbolic_attention",
|
| 30 |
"symbol_retrieval_kwargs": {
|
| 31 |
"d_model": 1024,
|
|
@@ -37,7 +32,5 @@
|
|
| 37 |
"shared_symbol_retriever": true,
|
| 38 |
"weight_tie_symbol_library": false
|
| 39 |
},
|
| 40 |
-
"torch_dtype": "float32",
|
| 41 |
-
"transformers_version": "4.39.3",
|
| 42 |
"vocab_size": 50304
|
| 43 |
-
}
|
|
|
|
| 1 |
{
|
| 2 |
"activation": "gelu",
|
|
|
|
|
|
|
|
|
|
| 3 |
"bias": false,
|
| 4 |
"d_model": 1024,
|
| 5 |
"dff": null,
|
|
|
|
| 9 |
"n_heads_sa": 8,
|
| 10 |
"n_layers": 24,
|
| 11 |
"norm_first": true,
|
|
|
|
| 12 |
"pos_enc_type": "RoPE",
|
| 13 |
"ra_kwargs": {
|
| 14 |
"n_kv_heads": 4,
|
|
|
|
| 21 |
"sa_kwargs": {
|
| 22 |
"n_kv_heads": 4
|
| 23 |
},
|
|
|
|
| 24 |
"symbol_retrieval": "symbolic_attention",
|
| 25 |
"symbol_retrieval_kwargs": {
|
| 26 |
"d_model": 1024,
|
|
|
|
| 32 |
"shared_symbol_retriever": true,
|
| 33 |
"weight_tie_symbol_library": false
|
| 34 |
},
|
|
|
|
|
|
|
| 35 |
"vocab_size": 50304
|
| 36 |
+
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:44e33f55dad515c97020e28a1975d29dc6e01628a3a35ca36026c4a831cfecaf
|
| 3 |
+
size 1377609760
|