rm attn_impl
Browse files- config.json +0 -4
config.json
CHANGED
|
@@ -34,7 +34,6 @@
|
|
| 34 |
"norm_eps": 1e-05,
|
| 35 |
"dropout": 0.0,
|
| 36 |
"rope_theta": 10000.0,
|
| 37 |
-
"attn_impl": "xformers",
|
| 38 |
"attn_bias_type": "local_block_causal",
|
| 39 |
"intermediate_size": 2048
|
| 40 |
},
|
|
@@ -56,7 +55,6 @@
|
|
| 56 |
"rope_type": "default"
|
| 57 |
},
|
| 58 |
"hidden_act": "silu",
|
| 59 |
-
"_attn_implementation": "sdpa",
|
| 60 |
"intermediate_size": 2816
|
| 61 |
},
|
| 62 |
"decoder_config": {
|
|
@@ -76,7 +74,6 @@
|
|
| 76 |
"rope_type": "default"
|
| 77 |
},
|
| 78 |
"hidden_act": "silu",
|
| 79 |
-
"_attn_implementation": "sdpa",
|
| 80 |
"intermediate_size": 2816
|
| 81 |
},
|
| 82 |
"global_config": {
|
|
@@ -92,7 +89,6 @@
|
|
| 92 |
"rope_type": "default"
|
| 93 |
},
|
| 94 |
"hidden_act": "silu",
|
| 95 |
-
"_attn_implementation": "sdpa",
|
| 96 |
"intermediate_size": 5632
|
| 97 |
},
|
| 98 |
"tie_word_embeddings": false
|
|
|
|
| 34 |
"norm_eps": 1e-05,
|
| 35 |
"dropout": 0.0,
|
| 36 |
"rope_theta": 10000.0,
|
|
|
|
| 37 |
"attn_bias_type": "local_block_causal",
|
| 38 |
"intermediate_size": 2048
|
| 39 |
},
|
|
|
|
| 55 |
"rope_type": "default"
|
| 56 |
},
|
| 57 |
"hidden_act": "silu",
|
|
|
|
| 58 |
"intermediate_size": 2816
|
| 59 |
},
|
| 60 |
"decoder_config": {
|
|
|
|
| 74 |
"rope_type": "default"
|
| 75 |
},
|
| 76 |
"hidden_act": "silu",
|
|
|
|
| 77 |
"intermediate_size": 2816
|
| 78 |
},
|
| 79 |
"global_config": {
|
|
|
|
| 89 |
"rope_type": "default"
|
| 90 |
},
|
| 91 |
"hidden_act": "silu",
|
|
|
|
| 92 |
"intermediate_size": 5632
|
| 93 |
},
|
| 94 |
"tie_word_embeddings": false
|