Upload folder using huggingface_hub
Browse files- config.json +3 -2
- special_tokens_map.json +1 -1
- tokenizer_config.json +1 -1
config.json
CHANGED
@@ -19,6 +19,7 @@
|
|
19 |
"vision_model",
|
20 |
"multi_modal_projector",
|
21 |
"lm_head",
|
|
|
22 |
"language_model.model.layers.0.feed_forward.router",
|
23 |
"language_model.model.layers.1.feed_forward.router",
|
24 |
"language_model.model.layers.2.feed_forward.router",
|
@@ -220,8 +221,8 @@
|
|
220 |
"pad_token_id": 200018,
|
221 |
"rms_norm_eps": 1e-05,
|
222 |
"rope_scaling": {
|
223 |
-
"factor":
|
224 |
-
"high_freq_factor":
|
225 |
"low_freq_factor": 1.0,
|
226 |
"original_max_position_embeddings": 8192,
|
227 |
"rope_type": "llama3"
|
|
|
19 |
"vision_model",
|
20 |
"multi_modal_projector",
|
21 |
"lm_head",
|
22 |
+
"vision_model",
|
23 |
"language_model.model.layers.0.feed_forward.router",
|
24 |
"language_model.model.layers.1.feed_forward.router",
|
25 |
"language_model.model.layers.2.feed_forward.router",
|
|
|
221 |
"pad_token_id": 200018,
|
222 |
"rms_norm_eps": 1e-05,
|
223 |
"rope_scaling": {
|
224 |
+
"factor": 16.0,
|
225 |
+
"high_freq_factor": 1.0,
|
226 |
"low_freq_factor": 1.0,
|
227 |
"original_max_position_embeddings": 8192,
|
228 |
"rope_type": "llama3"
|
special_tokens_map.json
CHANGED
@@ -14,7 +14,7 @@
|
|
14 |
"single_word": false
|
15 |
},
|
16 |
"pad_token": {
|
17 |
-
"content": "<|
|
18 |
"lstrip": false,
|
19 |
"normalized": false,
|
20 |
"rstrip": false,
|
|
|
14 |
"single_word": false
|
15 |
},
|
16 |
"pad_token": {
|
17 |
+
"content": "<|finetune_right_pad|>",
|
18 |
"lstrip": false,
|
19 |
"normalized": false,
|
20 |
"rstrip": false,
|
tokenizer_config.json
CHANGED
@@ -9091,7 +9091,7 @@
|
|
9091 |
"attention_mask"
|
9092 |
],
|
9093 |
"model_max_length": 10485760,
|
9094 |
-
"pad_token": "<|
|
9095 |
"processor_class": "Llama4Processor",
|
9096 |
"tokenizer_class": "PreTrainedTokenizer"
|
9097 |
}
|
|
|
9091 |
"attention_mask"
|
9092 |
],
|
9093 |
"model_max_length": 10485760,
|
9094 |
+
"pad_token": "<|finetune_right_pad|>",
|
9095 |
"processor_class": "Llama4Processor",
|
9096 |
"tokenizer_class": "PreTrainedTokenizer"
|
9097 |
}
|