Upload Idefics2ForConditionalGeneration
Browse files- config.json +9 -7
- generation_config.json +1 -0
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -23,17 +23,18 @@
|
|
| 23 |
"hidden_act": "silu",
|
| 24 |
"hidden_size": 16,
|
| 25 |
"initializer_range": 0.02,
|
| 26 |
-
"intermediate_size":
|
| 27 |
-
"max_position_embeddings":
|
| 28 |
"model_type": "mistral",
|
| 29 |
"num_attention_heads": 4,
|
| 30 |
"num_hidden_layers": 2,
|
| 31 |
"num_key_value_heads": 2,
|
| 32 |
-
"
|
|
|
|
| 33 |
"rope_theta": 10000.0,
|
| 34 |
"sliding_window": 4096,
|
| 35 |
"use_cache": true,
|
| 36 |
-
"vocab_size":
|
| 37 |
},
|
| 38 |
"tie_word_embeddings": false,
|
| 39 |
"torch_dtype": "float32",
|
|
@@ -43,14 +44,15 @@
|
|
| 43 |
"attention_dropout": 0.0,
|
| 44 |
"hidden_act": "gelu_pytorch_tanh",
|
| 45 |
"hidden_size": 16,
|
| 46 |
-
"image_size":
|
| 47 |
"initializer_range": 0.02,
|
| 48 |
-
"intermediate_size":
|
| 49 |
"layer_norm_eps": 1e-06,
|
| 50 |
"model_type": "idefics2_vision",
|
| 51 |
"num_attention_heads": 4,
|
| 52 |
"num_channels": 3,
|
| 53 |
"num_hidden_layers": 2,
|
| 54 |
-
"
|
|
|
|
| 55 |
}
|
| 56 |
}
|
|
|
|
| 23 |
"hidden_act": "silu",
|
| 24 |
"hidden_size": 16,
|
| 25 |
"initializer_range": 0.02,
|
| 26 |
+
"intermediate_size": 14336,
|
| 27 |
+
"max_position_embeddings": 32768,
|
| 28 |
"model_type": "mistral",
|
| 29 |
"num_attention_heads": 4,
|
| 30 |
"num_hidden_layers": 2,
|
| 31 |
"num_key_value_heads": 2,
|
| 32 |
+
"pad_token_id": 0,
|
| 33 |
+
"rms_norm_eps": 1e-05,
|
| 34 |
"rope_theta": 10000.0,
|
| 35 |
"sliding_window": 4096,
|
| 36 |
"use_cache": true,
|
| 37 |
+
"vocab_size": 32003
|
| 38 |
},
|
| 39 |
"tie_word_embeddings": false,
|
| 40 |
"torch_dtype": "float32",
|
|
|
|
| 44 |
"attention_dropout": 0.0,
|
| 45 |
"hidden_act": "gelu_pytorch_tanh",
|
| 46 |
"hidden_size": 16,
|
| 47 |
+
"image_size": 980,
|
| 48 |
"initializer_range": 0.02,
|
| 49 |
+
"intermediate_size": 4304,
|
| 50 |
"layer_norm_eps": 1e-06,
|
| 51 |
"model_type": "idefics2_vision",
|
| 52 |
"num_attention_heads": 4,
|
| 53 |
"num_channels": 3,
|
| 54 |
"num_hidden_layers": 2,
|
| 55 |
+
"num_key_value_heads": 2,
|
| 56 |
+
"patch_size": 14
|
| 57 |
}
|
| 58 |
}
|
generation_config.json
CHANGED
|
@@ -2,5 +2,6 @@
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
|
|
|
| 5 |
"transformers_version": "4.56.0.dev0"
|
| 6 |
}
|
|
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
"transformers_version": "4.56.0.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 sha256:4129e7380217a012ef101ff4aba519e4fde97091f87225ee2599a273e1a28d0c
|
| 3 |
+
size 14647936
|