Upload CohereForCausalLM
Browse files- config.json +5 -5
- generation_config.json +1 -1
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -5,13 +5,14 @@
|
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 5,
|
|
|
|
| 8 |
"eos_token_id": 255001,
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 8,
|
| 11 |
"initializer_range": 0.02,
|
| 12 |
"intermediate_size": 32,
|
| 13 |
"layer_norm_eps": 1e-05,
|
| 14 |
-
"logit_scale": 0.
|
| 15 |
"max_position_embeddings": 8192,
|
| 16 |
"model_type": "cohere",
|
| 17 |
"num_attention_heads": 4,
|
|
@@ -19,10 +20,9 @@
|
|
| 19 |
"num_key_value_heads": 2,
|
| 20 |
"pad_token_id": 0,
|
| 21 |
"rope_scaling": null,
|
| 22 |
-
"rope_theta": 10000
|
| 23 |
-
"
|
| 24 |
-
"transformers_version": "4.55.0.dev0",
|
| 25 |
"use_cache": true,
|
| 26 |
"use_qk_norm": false,
|
| 27 |
-
"vocab_size":
|
| 28 |
}
|
|
|
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 5,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
"eos_token_id": 255001,
|
| 10 |
"hidden_act": "silu",
|
| 11 |
"hidden_size": 8,
|
| 12 |
"initializer_range": 0.02,
|
| 13 |
"intermediate_size": 32,
|
| 14 |
"layer_norm_eps": 1e-05,
|
| 15 |
+
"logit_scale": 0.125,
|
| 16 |
"max_position_embeddings": 8192,
|
| 17 |
"model_type": "cohere",
|
| 18 |
"num_attention_heads": 4,
|
|
|
|
| 20 |
"num_key_value_heads": 2,
|
| 21 |
"pad_token_id": 0,
|
| 22 |
"rope_scaling": null,
|
| 23 |
+
"rope_theta": 10000,
|
| 24 |
+
"transformers_version": "4.57.0.dev0",
|
|
|
|
| 25 |
"use_cache": true,
|
| 26 |
"use_qk_norm": false,
|
| 27 |
+
"vocab_size": 256000
|
| 28 |
}
|
generation_config.json
CHANGED
|
@@ -3,5 +3,5 @@
|
|
| 3 |
"bos_token_id": 5,
|
| 4 |
"eos_token_id": 255001,
|
| 5 |
"pad_token_id": 0,
|
| 6 |
-
"transformers_version": "4.
|
| 7 |
}
|
|
|
|
| 3 |
"bos_token_id": 5,
|
| 4 |
"eos_token_id": 255001,
|
| 5 |
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.57.0.dev0"
|
| 7 |
}
|
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 sha256:700901ee456bbd5c45e819d695ed744b4541bab6e31955190822e02996a5bf1d
|
| 3 |
+
size 8201632
|