nintwentydo
commited on
Update config.json
Browse files- config.json +141 -11
config.json
CHANGED
@@ -1,13 +1,8 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"architectures": [
|
4 |
"LlavaForConditionalGeneration"
|
5 |
],
|
6 |
-
"ignore_index": -100,
|
7 |
-
"image_seq_length": 1,
|
8 |
-
"image_token_index": 10,
|
9 |
-
"model_type": "llava",
|
10 |
-
"projector_hidden_act": "gelu",
|
11 |
"quantization_config": {
|
12 |
"config_groups": {
|
13 |
"group_0": {
|
@@ -17,7 +12,7 @@
|
|
17 |
"dynamic": true,
|
18 |
"group_size": null,
|
19 |
"num_bits": 8,
|
20 |
-
"observer":
|
21 |
"observer_kwargs": {},
|
22 |
"strategy": "token",
|
23 |
"symmetric": true,
|
@@ -42,7 +37,7 @@
|
|
42 |
}
|
43 |
},
|
44 |
"format": "float-quantized",
|
45 |
-
"global_compression_ratio": 1.
|
46 |
"ignore": [
|
47 |
"vision_tower.transformer.layers.0.feed_forward.gate_proj",
|
48 |
"vision_tower.transformer.layers.0.feed_forward.up_proj",
|
@@ -231,31 +226,166 @@
|
|
231 |
"quant_method": "compressed-tensors",
|
232 |
"quantization_status": "compressed"
|
233 |
},
|
|
|
|
|
|
|
|
|
|
|
234 |
"text_config": {
|
235 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
236 |
"head_dim": 128,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
237 |
"intermediate_size": 14336,
|
238 |
"is_composition": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
239 |
"max_position_embeddings": 1024000,
|
|
|
240 |
"model_type": "mistral",
|
|
|
|
|
|
|
|
|
241 |
"num_hidden_layers": 40,
|
242 |
"num_key_value_heads": 8,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
243 |
"rms_norm_eps": 1e-05,
|
244 |
"rope_theta": 1000000000.0,
|
|
|
245 |
"sliding_window": null,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
246 |
"vocab_size": 131072
|
247 |
},
|
248 |
"torch_dtype": "bfloat16",
|
249 |
-
"transformers_version": "4.
|
250 |
"vision_config": {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
251 |
"head_dim": 64,
|
252 |
"hidden_act": "silu",
|
|
|
|
|
|
|
|
|
|
|
253 |
"image_size": 1024,
|
|
|
254 |
"is_composition": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
255 |
"model_type": "pixtral",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
256 |
"patch_size": 16,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
257 |
"rope_theta": 10000.0,
|
258 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
259 |
},
|
260 |
"vision_feature_layer": -1,
|
261 |
"vision_feature_select_strategy": "full"
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "",
|
3 |
"architectures": [
|
4 |
"LlavaForConditionalGeneration"
|
5 |
],
|
|
|
|
|
|
|
|
|
|
|
6 |
"quantization_config": {
|
7 |
"config_groups": {
|
8 |
"group_0": {
|
|
|
12 |
"dynamic": true,
|
13 |
"group_size": null,
|
14 |
"num_bits": 8,
|
15 |
+
"observer": "memoryless",
|
16 |
"observer_kwargs": {},
|
17 |
"strategy": "token",
|
18 |
"symmetric": true,
|
|
|
37 |
}
|
38 |
},
|
39 |
"format": "float-quantized",
|
40 |
+
"global_compression_ratio": 1.1688641443181655,
|
41 |
"ignore": [
|
42 |
"vision_tower.transformer.layers.0.feed_forward.gate_proj",
|
43 |
"vision_tower.transformer.layers.0.feed_forward.up_proj",
|
|
|
226 |
"quant_method": "compressed-tensors",
|
227 |
"quantization_status": "compressed"
|
228 |
},
|
229 |
+
"ignore_index": -100,
|
230 |
+
"image_seq_length": 1,
|
231 |
+
"image_token_index": 10,
|
232 |
+
"model_type": "llava",
|
233 |
+
"projector_hidden_act": "gelu",
|
234 |
"text_config": {
|
235 |
+
"_name_or_path": "",
|
236 |
+
"add_cross_attention": false,
|
237 |
+
"architectures": null,
|
238 |
+
"attention_dropout": 0.0,
|
239 |
+
"bad_words_ids": null,
|
240 |
+
"begin_suppress_tokens": null,
|
241 |
+
"bos_token_id": 1,
|
242 |
+
"chunk_size_feed_forward": 0,
|
243 |
+
"cross_attention_hidden_size": null,
|
244 |
+
"decoder_start_token_id": null,
|
245 |
+
"diversity_penalty": 0.0,
|
246 |
+
"do_sample": false,
|
247 |
+
"early_stopping": false,
|
248 |
+
"encoder_no_repeat_ngram_size": 0,
|
249 |
+
"eos_token_id": 2,
|
250 |
+
"exponential_decay_length_penalty": null,
|
251 |
+
"finetuning_task": null,
|
252 |
+
"forced_bos_token_id": null,
|
253 |
+
"forced_eos_token_id": null,
|
254 |
"head_dim": 128,
|
255 |
+
"hidden_act": "silu",
|
256 |
+
"hidden_size": 5120,
|
257 |
+
"id2label": {
|
258 |
+
"0": "LABEL_0",
|
259 |
+
"1": "LABEL_1"
|
260 |
+
},
|
261 |
+
"initializer_range": 0.02,
|
262 |
"intermediate_size": 14336,
|
263 |
"is_composition": true,
|
264 |
+
"is_decoder": false,
|
265 |
+
"is_encoder_decoder": false,
|
266 |
+
"label2id": {
|
267 |
+
"LABEL_0": 0,
|
268 |
+
"LABEL_1": 1
|
269 |
+
},
|
270 |
+
"length_penalty": 1.0,
|
271 |
+
"max_length": 20,
|
272 |
"max_position_embeddings": 1024000,
|
273 |
+
"min_length": 0,
|
274 |
"model_type": "mistral",
|
275 |
+
"no_repeat_ngram_size": 0,
|
276 |
+
"num_attention_heads": 32,
|
277 |
+
"num_beam_groups": 1,
|
278 |
+
"num_beams": 1,
|
279 |
"num_hidden_layers": 40,
|
280 |
"num_key_value_heads": 8,
|
281 |
+
"num_return_sequences": 1,
|
282 |
+
"output_attentions": false,
|
283 |
+
"output_hidden_states": false,
|
284 |
+
"output_scores": false,
|
285 |
+
"pad_token_id": null,
|
286 |
+
"prefix": null,
|
287 |
+
"problem_type": null,
|
288 |
+
"pruned_heads": {},
|
289 |
+
"remove_invalid_values": false,
|
290 |
+
"repetition_penalty": 1.0,
|
291 |
+
"return_dict": true,
|
292 |
+
"return_dict_in_generate": false,
|
293 |
"rms_norm_eps": 1e-05,
|
294 |
"rope_theta": 1000000000.0,
|
295 |
+
"sep_token_id": null,
|
296 |
"sliding_window": null,
|
297 |
+
"suppress_tokens": null,
|
298 |
+
"task_specific_params": null,
|
299 |
+
"temperature": 1.0,
|
300 |
+
"tf_legacy_loss": false,
|
301 |
+
"tie_encoder_decoder": false,
|
302 |
+
"tie_word_embeddings": false,
|
303 |
+
"tokenizer_class": null,
|
304 |
+
"top_k": 50,
|
305 |
+
"top_p": 1.0,
|
306 |
+
"torch_dtype": null,
|
307 |
+
"torchscript": false,
|
308 |
+
"typical_p": 1.0,
|
309 |
+
"use_bfloat16": false,
|
310 |
+
"use_cache": true,
|
311 |
"vocab_size": 131072
|
312 |
},
|
313 |
"torch_dtype": "bfloat16",
|
314 |
+
"transformers_version": "4.45.1",
|
315 |
"vision_config": {
|
316 |
+
"_name_or_path": "",
|
317 |
+
"add_cross_attention": false,
|
318 |
+
"architectures": null,
|
319 |
+
"attention_dropout": 0.0,
|
320 |
+
"bad_words_ids": null,
|
321 |
+
"begin_suppress_tokens": null,
|
322 |
+
"bos_token_id": null,
|
323 |
+
"chunk_size_feed_forward": 0,
|
324 |
+
"cross_attention_hidden_size": null,
|
325 |
+
"decoder_start_token_id": null,
|
326 |
+
"diversity_penalty": 0.0,
|
327 |
+
"do_sample": false,
|
328 |
+
"early_stopping": false,
|
329 |
+
"encoder_no_repeat_ngram_size": 0,
|
330 |
+
"eos_token_id": null,
|
331 |
+
"exponential_decay_length_penalty": null,
|
332 |
+
"finetuning_task": null,
|
333 |
+
"forced_bos_token_id": null,
|
334 |
+
"forced_eos_token_id": null,
|
335 |
"head_dim": 64,
|
336 |
"hidden_act": "silu",
|
337 |
+
"hidden_size": 1024,
|
338 |
+
"id2label": {
|
339 |
+
"0": "LABEL_0",
|
340 |
+
"1": "LABEL_1"
|
341 |
+
},
|
342 |
"image_size": 1024,
|
343 |
+
"intermediate_size": 4096,
|
344 |
"is_composition": true,
|
345 |
+
"is_decoder": false,
|
346 |
+
"is_encoder_decoder": false,
|
347 |
+
"label2id": {
|
348 |
+
"LABEL_0": 0,
|
349 |
+
"LABEL_1": 1
|
350 |
+
},
|
351 |
+
"length_penalty": 1.0,
|
352 |
+
"max_length": 20,
|
353 |
+
"min_length": 0,
|
354 |
"model_type": "pixtral",
|
355 |
+
"no_repeat_ngram_size": 0,
|
356 |
+
"num_attention_heads": 16,
|
357 |
+
"num_beam_groups": 1,
|
358 |
+
"num_beams": 1,
|
359 |
+
"num_channels": 3,
|
360 |
+
"num_hidden_layers": 24,
|
361 |
+
"num_return_sequences": 1,
|
362 |
+
"output_attentions": false,
|
363 |
+
"output_hidden_states": false,
|
364 |
+
"output_scores": false,
|
365 |
+
"pad_token_id": null,
|
366 |
"patch_size": 16,
|
367 |
+
"prefix": null,
|
368 |
+
"problem_type": null,
|
369 |
+
"pruned_heads": {},
|
370 |
+
"remove_invalid_values": false,
|
371 |
+
"repetition_penalty": 1.0,
|
372 |
+
"return_dict": true,
|
373 |
+
"return_dict_in_generate": false,
|
374 |
"rope_theta": 10000.0,
|
375 |
+
"sep_token_id": null,
|
376 |
+
"suppress_tokens": null,
|
377 |
+
"task_specific_params": null,
|
378 |
+
"temperature": 1.0,
|
379 |
+
"tf_legacy_loss": false,
|
380 |
+
"tie_encoder_decoder": false,
|
381 |
+
"tie_word_embeddings": false,
|
382 |
+
"tokenizer_class": null,
|
383 |
+
"top_k": 50,
|
384 |
+
"top_p": 1.0,
|
385 |
+
"torch_dtype": null,
|
386 |
+
"torchscript": false,
|
387 |
+
"typical_p": 1.0,
|
388 |
+
"use_bfloat16": false
|
389 |
},
|
390 |
"vision_feature_layer": -1,
|
391 |
"vision_feature_select_strategy": "full"
|