Upload LlavaNextForConditionalGeneration
Browse files- config.json +5 -3
- generation_config.json +1 -1
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -2,6 +2,7 @@
|
|
| 2 |
"architectures": [
|
| 3 |
"LlavaNextForConditionalGeneration"
|
| 4 |
],
|
|
|
|
| 5 |
"ignore_index": -100,
|
| 6 |
"image_grid_pinpoints": [
|
| 7 |
[
|
|
@@ -36,11 +37,13 @@
|
|
| 36 |
"MistralForCausalLM"
|
| 37 |
],
|
| 38 |
"attention_dropout": 0.0,
|
|
|
|
| 39 |
"head_dim": null,
|
| 40 |
"hidden_act": "silu",
|
| 41 |
"hidden_size": 16,
|
| 42 |
"initializer_range": 0.02,
|
| 43 |
"intermediate_size": 14336,
|
|
|
|
| 44 |
"max_position_embeddings": 32768,
|
| 45 |
"model_type": "mistral",
|
| 46 |
"num_attention_heads": 4,
|
|
@@ -49,16 +52,15 @@
|
|
| 49 |
"rms_norm_eps": 1e-05,
|
| 50 |
"rope_theta": 1000000.0,
|
| 51 |
"sliding_window": null,
|
| 52 |
-
"torch_dtype": "bfloat16",
|
| 53 |
"use_cache": true,
|
| 54 |
"vocab_size": 32064
|
| 55 |
},
|
| 56 |
"tie_word_embeddings": false,
|
| 57 |
-
"
|
| 58 |
-
"transformers_version": "4.56.0.dev0",
|
| 59 |
"use_image_newline_parameter": true,
|
| 60 |
"vision_config": {
|
| 61 |
"attention_dropout": 0.0,
|
|
|
|
| 62 |
"hidden_act": "quick_gelu",
|
| 63 |
"hidden_size": 16,
|
| 64 |
"image_size": 336,
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"LlavaNextForConditionalGeneration"
|
| 4 |
],
|
| 5 |
+
"dtype": "bfloat16",
|
| 6 |
"ignore_index": -100,
|
| 7 |
"image_grid_pinpoints": [
|
| 8 |
[
|
|
|
|
| 37 |
"MistralForCausalLM"
|
| 38 |
],
|
| 39 |
"attention_dropout": 0.0,
|
| 40 |
+
"dtype": "bfloat16",
|
| 41 |
"head_dim": null,
|
| 42 |
"hidden_act": "silu",
|
| 43 |
"hidden_size": 16,
|
| 44 |
"initializer_range": 0.02,
|
| 45 |
"intermediate_size": 14336,
|
| 46 |
+
"layer_types": null,
|
| 47 |
"max_position_embeddings": 32768,
|
| 48 |
"model_type": "mistral",
|
| 49 |
"num_attention_heads": 4,
|
|
|
|
| 52 |
"rms_norm_eps": 1e-05,
|
| 53 |
"rope_theta": 1000000.0,
|
| 54 |
"sliding_window": null,
|
|
|
|
| 55 |
"use_cache": true,
|
| 56 |
"vocab_size": 32064
|
| 57 |
},
|
| 58 |
"tie_word_embeddings": false,
|
| 59 |
+
"transformers_version": "4.57.0.dev0",
|
|
|
|
| 60 |
"use_image_newline_parameter": true,
|
| 61 |
"vision_config": {
|
| 62 |
"attention_dropout": 0.0,
|
| 63 |
+
"embed_dim": 32,
|
| 64 |
"hidden_act": "quick_gelu",
|
| 65 |
"hidden_size": 16,
|
| 66 |
"image_size": 336,
|
generation_config.json
CHANGED
|
@@ -2,5 +2,5 @@
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
-
"transformers_version": "4.
|
| 6 |
}
|
|
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.57.0.dev0"
|
| 6 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5399728
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:346f654b90cd0fafed9c3417b66251d23252047a55a994e5c699803f3d7b7317
|
| 3 |
size 5399728
|