Hugging Face's logo

TheStageAI
/
gemma-4-E2B-it

Image-Text-to-Text
MLX
GGUF
English
multilingual
gemma4
gemma
gemma-4
edge
on-device
apple-silicon
quantization
gptq
aqlm
ple
conversational
Model card Files
xet
 Community
quazim commited on
Commit
f0480c9
·
verified ·
1 Parent(s): f8598fe

Upload folder using huggingface_hub

Browse files
Files changed (6) hide show
  1. .gitattributes +1 -0
  2. config.json +615 -0
  3. model.safetensors +3 -0
  4. ple.safetensors +3 -0
  5. tokenizer.json +3 -0
  6. tokenizer_config.json +74 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
config.json ADDED
@@ -0,0 +1,615 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma4ForConditionalGeneration"
4
+ ],
5
+ "audio_config": null,
6
+ "audio_token_id": 258881,
7
+ "boa_token_id": 256000,
8
+ "boi_token_id": 255999,
9
+ "dtype": "bfloat16",
10
+ "eoa_token_id": 258883,
11
+ "eoa_token_index": 258883,
12
+ "eoi_token_id": 258882,
13
+ "eos_token_id": [
14
+ 1,
15
+ 106
16
+ ],
17
+ "image_token_id": 258880,
18
+ "initializer_range": 0.02,
19
+ "model_type": "gemma4",
20
+ "text_config": {
21
+ "attention_bias": false,
22
+ "attention_dropout": 0.0,
23
+ "attention_k_eq_v": false,
24
+ "bos_token_id": 2,
25
+ "dtype": "bfloat16",
26
+ "enable_moe_block": false,
27
+ "eos_token_id": 1,
28
+ "expert_intermediate_size": null,
29
+ "final_logit_softcapping": 30.0,
30
+ "global_head_dim": 512,
31
+ "head_dim": 256,
32
+ "hidden_activation": "gelu_pytorch_tanh",
33
+ "hidden_size": 1536,
34
+ "hidden_size_per_layer_input": 256,
35
+ "initializer_range": 0.02,
36
+ "intermediate_size": 6144,
37
+ "layer_types": [
38
+ "sliding_attention",
39
+ "sliding_attention",
40
+ "sliding_attention",
41
+ "sliding_attention",
42
+ "full_attention",
43
+ "sliding_attention",
44
+ "sliding_attention",
45
+ "sliding_attention",
46
+ "sliding_attention",
47
+ "full_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "sliding_attention",
51
+ "sliding_attention",
52
+ "full_attention",
53
+ "sliding_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "sliding_attention",
57
+ "full_attention",
58
+ "sliding_attention",
59
+ "sliding_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "full_attention",
63
+ "sliding_attention",
64
+ "sliding_attention",
65
+ "sliding_attention",
66
+ "sliding_attention",
67
+ "full_attention",
68
+ "sliding_attention",
69
+ "sliding_attention",
70
+ "sliding_attention",
71
+ "sliding_attention",
72
+ "full_attention"
73
+ ],
74
+ "max_position_embeddings": 131072,
75
+ "model_type": "gemma4_text",
76
+ "num_attention_heads": 8,
77
+ "num_experts": null,
78
+ "num_global_key_value_heads": null,
79
+ "num_hidden_layers": 35,
80
+ "num_key_value_heads": 1,
81
+ "num_kv_shared_layers": 20,
82
+ "pad_token_id": 0,
83
+ "rms_norm_eps": 1e-06,
84
+ "rope_parameters": {
85
+ "full_attention": {
86
+ "partial_rotary_factor": 0.25,
87
+ "rope_theta": 1000000.0,
88
+ "rope_type": "proportional"
89
+ },
90
+ "sliding_attention": {
91
+ "rope_theta": 10000.0,
92
+ "rope_type": "default"
93
+ }
94
+ },
95
+ "sliding_window": 512,
96
+ "tie_word_embeddings": true,
97
+ "top_k_experts": null,
98
+ "use_bidirectional_attention": null,
99
+ "use_cache": true,
100
+ "use_double_wide_mlp": true,
101
+ "vocab_size": 262144,
102
+ "vocab_size_per_layer_input": 262144
103
+ },
104
+ "tie_word_embeddings": true,
105
+ "transformers_version": "5.5.0.dev0",
106
+ "video_token_id": 258884,
107
+ "vision_config": {
108
+ "_name_or_path": "",
109
+ "architectures": null,
110
+ "attention_bias": false,
111
+ "attention_dropout": 0.0,
112
+ "chunk_size_feed_forward": 0,
113
+ "default_output_length": 280,
114
+ "dtype": "bfloat16",
115
+ "global_head_dim": 64,
116
+ "head_dim": 64,
117
+ "hidden_activation": "gelu_pytorch_tanh",
118
+ "hidden_size": 768,
119
+ "id2label": {
120
+ "0": "LABEL_0",
121
+ "1": "LABEL_1"
122
+ },
123
+ "initializer_range": 0.02,
124
+ "intermediate_size": 3072,
125
+ "is_encoder_decoder": false,
126
+ "label2id": {
127
+ "LABEL_0": 0,
128
+ "LABEL_1": 1
129
+ },
130
+ "max_position_embeddings": 131072,
131
+ "model_type": "gemma4_vision",
132
+ "num_attention_heads": 12,
133
+ "num_hidden_layers": 16,
134
+ "num_key_value_heads": 12,
135
+ "output_attentions": false,
136
+ "output_hidden_states": false,
137
+ "patch_size": 16,
138
+ "pooling_kernel_size": 3,
139
+ "position_embedding_size": 10240,
140
+ "problem_type": null,
141
+ "return_dict": true,
142
+ "rms_norm_eps": 1e-06,
143
+ "rope_parameters": {
144
+ "rope_theta": 100.0,
145
+ "rope_type": "default"
146
+ },
147
+ "standardize": false,
148
+ "use_clipped_linears": true
149
+ },
150
+ "vision_soft_tokens_per_image": 280,
151
+ "quantization": {
152
+ "group_size": 64,
153
+ "bits": 4,
154
+ "language_model.model.embed_tokens": {
155
+ "bits": 3,
156
+ "group_size": 64
157
+ },
158
+ "language_model.model.layers.0.mlp.down_proj": {
159
+ "bits": 3,
160
+ "group_size": 64
161
+ },
162
+ "language_model.model.layers.0.mlp.gate_proj": {
163
+ "bits": 3,
164
+ "group_size": 64
165
+ },
166
+ "language_model.model.layers.0.mlp.up_proj": {
167
+ "bits": 3,
168
+ "group_size": 64
169
+ },
170
+ "language_model.model.layers.0.self_attn.k_proj": {
171
+ "bits": 3,
172
+ "group_size": 32
173
+ },
174
+ "language_model.model.layers.0.self_attn.v_proj": {
175
+ "bits": 3,
176
+ "group_size": 32
177
+ },
178
+ "language_model.model.layers.1.mlp.down_proj": {
179
+ "bits": 3,
180
+ "group_size": 64
181
+ },
182
+ "language_model.model.layers.1.mlp.gate_proj": {
183
+ "bits": 3,
184
+ "group_size": 64
185
+ },
186
+ "language_model.model.layers.1.mlp.up_proj": {
187
+ "bits": 3,
188
+ "group_size": 64
189
+ },
190
+ "language_model.model.layers.1.self_attn.k_proj": {
191
+ "bits": 3,
192
+ "group_size": 32
193
+ },
194
+ "language_model.model.layers.1.self_attn.v_proj": {
195
+ "bits": 3,
196
+ "group_size": 32
197
+ },
198
+ "language_model.model.layers.10.mlp.gate_proj": {
199
+ "bits": 3,
200
+ "group_size": 32
201
+ },
202
+ "language_model.model.layers.10.self_attn.k_proj": {
203
+ "bits": 3,
204
+ "group_size": 32
205
+ },
206
+ "language_model.model.layers.10.self_attn.v_proj": {
207
+ "bits": 3,
208
+ "group_size": 32
209
+ },
210
+ "language_model.model.layers.11.self_attn.k_proj": {
211
+ "bits": 3,
212
+ "group_size": 32
213
+ },
214
+ "language_model.model.layers.11.self_attn.v_proj": {
215
+ "bits": 3,
216
+ "group_size": 32
217
+ },
218
+ "language_model.model.layers.13.self_attn.o_proj": {
219
+ "bits": 3,
220
+ "group_size": 64
221
+ },
222
+ "language_model.model.layers.13.self_attn.q_proj": {
223
+ "bits": 3,
224
+ "group_size": 64
225
+ },
226
+ "language_model.model.layers.13.self_attn.v_proj": {
227
+ "bits": 3,
228
+ "group_size": 32
229
+ },
230
+ "language_model.model.layers.14.mlp.up_proj": {
231
+ "bits": 3,
232
+ "group_size": 32
233
+ },
234
+ "language_model.model.layers.15.self_attn.o_proj": {
235
+ "bits": 3,
236
+ "group_size": 32
237
+ },
238
+ "language_model.model.layers.15.self_attn.q_proj": {
239
+ "bits": 3,
240
+ "group_size": 64
241
+ },
242
+ "language_model.model.layers.16.mlp.down_proj": {
243
+ "bits": 3,
244
+ "group_size": 32
245
+ },
246
+ "language_model.model.layers.16.mlp.gate_proj": {
247
+ "bits": 3,
248
+ "group_size": 64
249
+ },
250
+ "language_model.model.layers.16.self_attn.o_proj": {
251
+ "bits": 3,
252
+ "group_size": 64
253
+ },
254
+ "language_model.model.layers.16.self_attn.q_proj": {
255
+ "bits": 3,
256
+ "group_size": 64
257
+ },
258
+ "language_model.model.layers.17.mlp.gate_proj": {
259
+ "bits": 3,
260
+ "group_size": 64
261
+ },
262
+ "language_model.model.layers.17.self_attn.o_proj": {
263
+ "bits": 3,
264
+ "group_size": 64
265
+ },
266
+ "language_model.model.layers.17.self_attn.q_proj": {
267
+ "bits": 3,
268
+ "group_size": 64
269
+ },
270
+ "language_model.model.layers.18.self_attn.o_proj": {
271
+ "bits": 3,
272
+ "group_size": 64
273
+ },
274
+ "language_model.model.layers.18.self_attn.q_proj": {
275
+ "bits": 3,
276
+ "group_size": 64
277
+ },
278
+ "language_model.model.layers.19.mlp.down_proj": {
279
+ "bits": 3,
280
+ "group_size": 64
281
+ },
282
+ "language_model.model.layers.19.mlp.gate_proj": {
283
+ "bits": 3,
284
+ "group_size": 64
285
+ },
286
+ "language_model.model.layers.19.mlp.up_proj": {
287
+ "bits": 3,
288
+ "group_size": 32
289
+ },
290
+ "language_model.model.layers.19.self_attn.o_proj": {
291
+ "bits": 3,
292
+ "group_size": 32
293
+ },
294
+ "language_model.model.layers.19.self_attn.q_proj": {
295
+ "bits": 3,
296
+ "group_size": 64
297
+ },
298
+ "language_model.model.layers.2.mlp.down_proj": {
299
+ "bits": 3,
300
+ "group_size": 64
301
+ },
302
+ "language_model.model.layers.2.mlp.gate_proj": {
303
+ "bits": 3,
304
+ "group_size": 64
305
+ },
306
+ "language_model.model.layers.2.mlp.up_proj": {
307
+ "bits": 3,
308
+ "group_size": 64
309
+ },
310
+ "language_model.model.layers.2.self_attn.k_proj": {
311
+ "bits": 3,
312
+ "group_size": 32
313
+ },
314
+ "language_model.model.layers.2.self_attn.v_proj": {
315
+ "bits": 3,
316
+ "group_size": 32
317
+ },
318
+ "language_model.model.layers.21.self_attn.o_proj": {
319
+ "bits": 2,
320
+ "group_size": 32
321
+ },
322
+ "language_model.model.layers.21.self_attn.q_proj": {
323
+ "bits": 2,
324
+ "group_size": 32
325
+ },
326
+ "language_model.model.layers.22.mlp.gate_proj": {
327
+ "bits": 3,
328
+ "group_size": 64
329
+ },
330
+ "language_model.model.layers.23.mlp.up_proj": {
331
+ "bits": 3,
332
+ "group_size": 32
333
+ },
334
+ "language_model.model.layers.24.self_attn.o_proj": {
335
+ "bits": 3,
336
+ "group_size": 32
337
+ },
338
+ "language_model.model.layers.24.self_attn.q_proj": {
339
+ "bits": 3,
340
+ "group_size": 64
341
+ },
342
+ "language_model.model.layers.25.mlp.up_proj": {
343
+ "bits": 3,
344
+ "group_size": 32
345
+ },
346
+ "language_model.model.layers.25.self_attn.o_proj": {
347
+ "bits": 3,
348
+ "group_size": 64
349
+ },
350
+ "language_model.model.layers.25.self_attn.q_proj": {
351
+ "bits": 3,
352
+ "group_size": 64
353
+ },
354
+ "language_model.model.layers.26.mlp.down_proj": {
355
+ "bits": 3,
356
+ "group_size": 32
357
+ },
358
+ "language_model.model.layers.26.mlp.up_proj": {
359
+ "bits": 3,
360
+ "group_size": 64
361
+ },
362
+ "language_model.model.layers.26.self_attn.q_proj": {
363
+ "bits": 3,
364
+ "group_size": 64
365
+ },
366
+ "language_model.model.layers.27.mlp.down_proj": {
367
+ "bits": 3,
368
+ "group_size": 64
369
+ },
370
+ "language_model.model.layers.27.mlp.gate_proj": {
371
+ "bits": 3,
372
+ "group_size": 64
373
+ },
374
+ "language_model.model.layers.27.mlp.up_proj": {
375
+ "bits": 3,
376
+ "group_size": 64
377
+ },
378
+ "language_model.model.layers.27.self_attn.o_proj": {
379
+ "bits": 3,
380
+ "group_size": 64
381
+ },
382
+ "language_model.model.layers.27.self_attn.q_proj": {
383
+ "bits": 2,
384
+ "group_size": 32
385
+ },
386
+ "language_model.model.layers.28.mlp.down_proj": {
387
+ "bits": 3,
388
+ "group_size": 64
389
+ },
390
+ "language_model.model.layers.28.mlp.gate_proj": {
391
+ "bits": 3,
392
+ "group_size": 64
393
+ },
394
+ "language_model.model.layers.28.mlp.up_proj": {
395
+ "bits": 3,
396
+ "group_size": 64
397
+ },
398
+ "language_model.model.layers.28.self_attn.o_proj": {
399
+ "bits": 3,
400
+ "group_size": 64
401
+ },
402
+ "language_model.model.layers.28.self_attn.q_proj": {
403
+ "bits": 3,
404
+ "group_size": 64
405
+ },
406
+ "language_model.model.layers.29.mlp.down_proj": {
407
+ "bits": 3,
408
+ "group_size": 64
409
+ },
410
+ "language_model.model.layers.29.mlp.gate_proj": {
411
+ "bits": 3,
412
+ "group_size": 64
413
+ },
414
+ "language_model.model.layers.29.mlp.up_proj": {
415
+ "bits": 3,
416
+ "group_size": 64
417
+ },
418
+ "language_model.model.layers.3.mlp.down_proj": {
419
+ "bits": 3,
420
+ "group_size": 64
421
+ },
422
+ "language_model.model.layers.3.mlp.gate_proj": {
423
+ "bits": 3,
424
+ "group_size": 64
425
+ },
426
+ "language_model.model.layers.3.self_attn.k_proj": {
427
+ "bits": 3,
428
+ "group_size": 32
429
+ },
430
+ "language_model.model.layers.3.self_attn.o_proj": {
431
+ "bits": 3,
432
+ "group_size": 64
433
+ },
434
+ "language_model.model.layers.3.self_attn.q_proj": {
435
+ "bits": 3,
436
+ "group_size": 64
437
+ },
438
+ "language_model.model.layers.3.self_attn.v_proj": {
439
+ "bits": 3,
440
+ "group_size": 32
441
+ },
442
+ "language_model.model.layers.30.mlp.down_proj": {
443
+ "bits": 3,
444
+ "group_size": 64
445
+ },
446
+ "language_model.model.layers.30.mlp.gate_proj": {
447
+ "bits": 3,
448
+ "group_size": 64
449
+ },
450
+ "language_model.model.layers.30.mlp.up_proj": {
451
+ "bits": 3,
452
+ "group_size": 64
453
+ },
454
+ "language_model.model.layers.30.self_attn.o_proj": {
455
+ "bits": 3,
456
+ "group_size": 64
457
+ },
458
+ "language_model.model.layers.30.self_attn.q_proj": {
459
+ "bits": 2,
460
+ "group_size": 32
461
+ },
462
+ "language_model.model.layers.31.mlp.down_proj": {
463
+ "bits": 3,
464
+ "group_size": 64
465
+ },
466
+ "language_model.model.layers.31.mlp.gate_proj": {
467
+ "bits": 3,
468
+ "group_size": 64
469
+ },
470
+ "language_model.model.layers.31.mlp.up_proj": {
471
+ "bits": 3,
472
+ "group_size": 64
473
+ },
474
+ "language_model.model.layers.31.self_attn.o_proj": {
475
+ "bits": 2,
476
+ "group_size": 32
477
+ },
478
+ "language_model.model.layers.31.self_attn.q_proj": {
479
+ "bits": 3,
480
+ "group_size": 64
481
+ },
482
+ "language_model.model.layers.32.mlp.down_proj": {
483
+ "bits": 3,
484
+ "group_size": 64
485
+ },
486
+ "language_model.model.layers.32.mlp.gate_proj": {
487
+ "bits": 3,
488
+ "group_size": 64
489
+ },
490
+ "language_model.model.layers.32.mlp.up_proj": {
491
+ "bits": 3,
492
+ "group_size": 64
493
+ },
494
+ "language_model.model.layers.32.self_attn.o_proj": {
495
+ "bits": 2,
496
+ "group_size": 32
497
+ },
498
+ "language_model.model.layers.32.self_attn.q_proj": {
499
+ "bits": 2,
500
+ "group_size": 32
501
+ },
502
+ "language_model.model.layers.33.mlp.gate_proj": {
503
+ "bits": 3,
504
+ "group_size": 64
505
+ },
506
+ "language_model.model.layers.33.mlp.up_proj": {
507
+ "bits": 3,
508
+ "group_size": 64
509
+ },
510
+ "language_model.model.layers.33.self_attn.o_proj": {
511
+ "bits": 3,
512
+ "group_size": 32
513
+ },
514
+ "language_model.model.layers.33.self_attn.q_proj": {
515
+ "bits": 3,
516
+ "group_size": 64
517
+ },
518
+ "language_model.model.layers.34.mlp.down_proj": {
519
+ "bits": 3,
520
+ "group_size": 64
521
+ },
522
+ "language_model.model.layers.34.mlp.gate_proj": {
523
+ "bits": 3,
524
+ "group_size": 64
525
+ },
526
+ "language_model.model.layers.34.mlp.up_proj": {
527
+ "bits": 3,
528
+ "group_size": 64
529
+ },
530
+ "language_model.model.layers.34.self_attn.o_proj": {
531
+ "bits": 3,
532
+ "group_size": 64
533
+ },
534
+ "language_model.model.layers.34.self_attn.q_proj": {
535
+ "bits": 3,
536
+ "group_size": 64
537
+ },
538
+ "language_model.model.layers.4.mlp.gate_proj": {
539
+ "bits": 3,
540
+ "group_size": 64
541
+ },
542
+ "language_model.model.layers.4.mlp.up_proj": {
543
+ "bits": 3,
544
+ "group_size": 64
545
+ },
546
+ "language_model.model.layers.4.self_attn.o_proj": {
547
+ "bits": 3,
548
+ "group_size": 64
549
+ },
550
+ "language_model.model.layers.4.self_attn.q_proj": {
551
+ "bits": 3,
552
+ "group_size": 64
553
+ },
554
+ "language_model.model.layers.5.mlp.up_proj": {
555
+ "bits": 3,
556
+ "group_size": 64
557
+ },
558
+ "language_model.model.layers.5.self_attn.k_proj": {
559
+ "bits": 3,
560
+ "group_size": 32
561
+ },
562
+ "language_model.model.layers.5.self_attn.v_proj": {
563
+ "bits": 3,
564
+ "group_size": 32
565
+ },
566
+ "language_model.model.layers.6.mlp.gate_proj": {
567
+ "bits": 3,
568
+ "group_size": 64
569
+ },
570
+ "language_model.model.layers.6.mlp.up_proj": {
571
+ "bits": 3,
572
+ "group_size": 64
573
+ },
574
+ "language_model.model.layers.6.self_attn.k_proj": {
575
+ "bits": 3,
576
+ "group_size": 32
577
+ },
578
+ "language_model.model.layers.6.self_attn.v_proj": {
579
+ "bits": 3,
580
+ "group_size": 32
581
+ },
582
+ "language_model.model.layers.7.mlp.down_proj": {
583
+ "bits": 3,
584
+ "group_size": 64
585
+ },
586
+ "language_model.model.layers.7.mlp.up_proj": {
587
+ "bits": 3,
588
+ "group_size": 64
589
+ },
590
+ "language_model.model.layers.7.self_attn.k_proj": {
591
+ "bits": 3,
592
+ "group_size": 32
593
+ },
594
+ "language_model.model.layers.7.self_attn.v_proj": {
595
+ "bits": 3,
596
+ "group_size": 32
597
+ },
598
+ "language_model.model.layers.8.mlp.gate_proj": {
599
+ "bits": 3,
600
+ "group_size": 64
601
+ },
602
+ "language_model.model.layers.8.self_attn.k_proj": {
603
+ "bits": 3,
604
+ "group_size": 32
605
+ },
606
+ "language_model.model.layers.8.self_attn.v_proj": {
607
+ "bits": 3,
608
+ "group_size": 32
609
+ },
610
+ "language_model.model.layers.9.self_attn.v_proj": {
611
+ "bits": 3,
612
+ "group_size": 32
613
+ }
614
+ }
615
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fabdcd8b1e5066000e557746a3560f42ce84afceb37e068dc8f93bd460a23c5b
3
+ size 1325634644
ple.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31ce9837c1b4211552ba97a1dbbba00df0be78ee580620ad550e80659cfca66e
3
+ size 259195352
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc8d3a0ce36466ccc1278bf987df5f71db1719b9ca6b4118264f45cb627bfe0f
3
+ size 32169626
tokenizer_config.json ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audio_token": "<|audio|>",
3
+ "backend": "tokenizers",
4
+ "boa_token": "<|audio>",
5
+ "boi_token": "<|image>",
6
+ "bos_token": "<bos>",
7
+ "eoa_token": "<audio|>",
8
+ "eoc_token": "<channel|>",
9
+ "eoi_token": "<image|>",
10
+ "eos_token": "<eos>",
11
+ "eot_token": "<turn|>",
12
+ "escape_token": "<|\"|>",
13
+ "etc_token": "<tool_call|>",
14
+ "etd_token": "<tool|>",
15
+ "etr_token": "<tool_response|>",
16
+ "extra_special_tokens": [
17
+ "<|video|>"
18
+ ],
19
+ "image_token": "<|image|>",
20
+ "mask_token": "<mask>",
21
+ "model_max_length": 1000000000000000019884624838656,
22
+ "pad_token": "<pad>",
23
+ "padding_side": "left",
24
+ "processor_class": "Gemma4Processor",
25
+ "response_schema": {
26
+ "type": "object",
27
+ "properties": {
28
+ "role": {
29
+ "const": "assistant"
30
+ },
31
+ "thinking": {
32
+ "type": "string"
33
+ },
34
+ "content": {
35
+ "type": "string"
36
+ },
37
+ "tool_calls": {
38
+ "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>",
39
+ "type": "array",
40
+ "items": {
41
+ "type": "object",
42
+ "properties": {
43
+ "type": {
44
+ "const": "function"
45
+ },
46
+ "function": {
47
+ "type": "object",
48
+ "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})",
49
+ "properties": {
50
+ "name": {
51
+ "type": "string"
52
+ },
53
+ "arguments": {
54
+ "type": "object",
55
+ "x-parser": "gemma4-tool-call",
56
+ "additionalProperties": {}
57
+ }
58
+ }
59
+ }
60
+ }
61
+ }
62
+ }
63
+ },
64
+ "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
65
+ },
66
+ "soc_token": "<|channel>",
67
+ "sot_token": "<|turn>",
68
+ "stc_token": "<|tool_call>",
69
+ "std_token": "<|tool>",
70
+ "str_token": "<|tool_response>",
71
+ "think_token": "<|think|>",
72
+ "tokenizer_class": "GemmaTokenizer",
73
+ "unk_token": "<unk>"
74
+ }