Upload config.json with huggingface_hub
Browse files- config.json +8 -6
config.json
CHANGED
|
@@ -2,19 +2,21 @@
|
|
| 2 |
"architectures": [
|
| 3 |
"SDARForCausalLM"
|
| 4 |
],
|
| 5 |
-
"attention_bias": false,
|
| 6 |
-
"attention_dropout": 0.0,
|
| 7 |
"auto_map": {
|
| 8 |
"AutoConfig": "configuration_sdar.SDARConfig",
|
| 9 |
"AutoModel": "modeling_sdar.SDARModel",
|
| 10 |
-
"AutoModelForCausalLM": "modeling_sdar.SDARForCausalLM"
|
| 11 |
-
"AutoTokenizer": "tokenization_qwen2.Qwen2Tokenizer"
|
| 12 |
},
|
|
|
|
|
|
|
|
|
|
| 13 |
"bos_token_id": 151643,
|
| 14 |
"debug": false,
|
| 15 |
"eos_token_id": 151643,
|
|
|
|
|
|
|
| 16 |
"ep_size": 1,
|
| 17 |
-
"fuse_cross_entropy":
|
| 18 |
"head_dim": 128,
|
| 19 |
"hidden_act": "silu",
|
| 20 |
"hidden_size": 4096,
|
|
@@ -39,4 +41,4 @@
|
|
| 39 |
"use_deepep": false,
|
| 40 |
"use_sliding_window": false,
|
| 41 |
"vocab_size": 151936
|
| 42 |
-
}
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"SDARForCausalLM"
|
| 4 |
],
|
|
|
|
|
|
|
| 5 |
"auto_map": {
|
| 6 |
"AutoConfig": "configuration_sdar.SDARConfig",
|
| 7 |
"AutoModel": "modeling_sdar.SDARModel",
|
| 8 |
+
"AutoModelForCausalLM": "modeling_sdar.SDARForCausalLM"
|
|
|
|
| 9 |
},
|
| 10 |
+
"attention_bias": false,
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"attn_implementation": "flex_attention",
|
| 13 |
"bos_token_id": 151643,
|
| 14 |
"debug": false,
|
| 15 |
"eos_token_id": 151643,
|
| 16 |
+
"block_size": 4,
|
| 17 |
+
"mask_token_id": 151669,
|
| 18 |
"ep_size": 1,
|
| 19 |
+
"fuse_cross_entropy": true,
|
| 20 |
"head_dim": 128,
|
| 21 |
"hidden_act": "silu",
|
| 22 |
"hidden_size": 4096,
|
|
|
|
| 41 |
"use_deepep": false,
|
| 42 |
"use_sliding_window": false,
|
| 43 |
"vocab_size": 151936
|
| 44 |
+
}
|