Hugging Face's logo

onnx-community
/
LFM2-VL-3B-ONNX

ONNX
lfm2_vl
Model card Files
xet
 Community
Xenova HF Staff commited on
Commit
c9ac3e2
·
verified ·
1 Parent(s): 70bc695

Upload folder using huggingface_hub

Browse files
Files changed (26) hide show
  1. .gitattributes +13 -0
  2. chat_template.jinja +4 -0
  3. config.json +126 -0
  4. generation_config.json +7 -0
  5. onnx/decoder_model_merged.onnx +3 -0
  6. onnx/decoder_model_merged.onnx_data +3 -0
  7. onnx/decoder_model_merged.onnx_data_1 +3 -0
  8. onnx/decoder_model_merged.onnx_data_2 +3 -0
  9. onnx/decoder_model_merged.onnx_data_3 +3 -0
  10. onnx/decoder_model_merged.onnx_data_4 +3 -0
  11. onnx/decoder_model_merged.onnx_data_5 +3 -0
  12. onnx/decoder_model_merged_fp16.onnx +3 -0
  13. onnx/decoder_model_merged_fp16.onnx_data +3 -0
  14. onnx/decoder_model_merged_fp16.onnx_data_1 +3 -0
  15. onnx/decoder_model_merged_fp16.onnx_data_2 +3 -0
  16. onnx/embed_tokens.onnx +3 -0
  17. onnx/embed_tokens.onnx_data +3 -0
  18. onnx/embed_tokens_fp16.onnx +3 -0
  19. onnx/embed_tokens_fp16.onnx_data +3 -0
  20. onnx/vision_encoder.onnx +3 -0
  21. onnx/vision_encoder.onnx_data +3 -0
  22. onnx/vision_encoder_fp16.onnx +3 -0
  23. onnx/vision_encoder_fp16.onnx_data +3 -0
  24. processor_config.json +51 -0
  25. tokenizer.json +0 -0
  26. tokenizer_config.json +29 -0
.gitattributes CHANGED
@@ -33,3 +33,16 @@ 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
+ onnx/decoder_model_merged.onnx_data filter=lfs diff=lfs merge=lfs -text
37
+ onnx/decoder_model_merged.onnx_data_1 filter=lfs diff=lfs merge=lfs -text
38
+ onnx/decoder_model_merged.onnx_data_2 filter=lfs diff=lfs merge=lfs -text
39
+ onnx/decoder_model_merged.onnx_data_3 filter=lfs diff=lfs merge=lfs -text
40
+ onnx/decoder_model_merged.onnx_data_4 filter=lfs diff=lfs merge=lfs -text
41
+ onnx/decoder_model_merged.onnx_data_5 filter=lfs diff=lfs merge=lfs -text
42
+ onnx/decoder_model_merged_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
43
+ onnx/decoder_model_merged_fp16.onnx_data_1 filter=lfs diff=lfs merge=lfs -text
44
+ onnx/decoder_model_merged_fp16.onnx_data_2 filter=lfs diff=lfs merge=lfs -text
45
+ onnx/embed_tokens.onnx_data filter=lfs diff=lfs merge=lfs -text
46
+ onnx/embed_tokens_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
47
+ onnx/vision_encoder.onnx_data filter=lfs diff=lfs merge=lfs -text
48
+ onnx/vision_encoder_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
chat_template.jinja ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {{bos_token}}{% for message in messages %}{{'<|im_start|>' + message['role'] + '
2
+ '}}{% if message['content'] is string %}{{ message['content'] }}{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' %}{{ '<image>' }}{% elif content['type'] == 'text' %}{{ content['text'] }}{% endif %}{% endfor %}{% endif %}{{'<|im_end|>
3
+ '}}{% endfor %}{% if add_generation_prompt %}{{'<|im_start|>assistant
4
+ ' }}{% endif %}
config.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Lfm2VlForConditionalGeneration"
4
+ ],
5
+ "do_image_splitting": true,
6
+ "downsample_factor": 2,
7
+ "dtype": "bfloat16",
8
+ "encoder_patch_size": 16,
9
+ "image_token_id": 396,
10
+ "max_image_tokens": 256,
11
+ "max_pixels_tolerance": 2.0,
12
+ "max_tiles": 10,
13
+ "min_image_tokens": 64,
14
+ "min_tiles": 2,
15
+ "model_type": "lfm2_vl",
16
+ "projector_bias": true,
17
+ "projector_hidden_act": "gelu",
18
+ "projector_hidden_size": 2560,
19
+ "text_config": {
20
+ "_name_or_path": "LiquidAI/LFM2-2.6B",
21
+ "architectures": [
22
+ "Lfm2ForCausalLM"
23
+ ],
24
+ "block_auto_adjust_ff_dim": false,
25
+ "block_dim": 2048,
26
+ "block_ff_dim": 10752,
27
+ "block_ffn_dim_multiplier": 1.0,
28
+ "block_mlp_init_scale": 1.0,
29
+ "block_multiple_of": 256,
30
+ "block_norm_eps": 1e-05,
31
+ "block_out_init_scale": 1.0,
32
+ "block_use_swiglu": true,
33
+ "block_use_xavier_init": true,
34
+ "conv_L_cache": 3,
35
+ "conv_bias": false,
36
+ "conv_dim": 2048,
37
+ "conv_dim_out": 2048,
38
+ "conv_use_xavier_init": true,
39
+ "dtype": "bfloat16",
40
+ "eos_token_id": 7,
41
+ "hidden_size": 2048,
42
+ "initializer_range": 0.02,
43
+ "intermediate_size": 10752,
44
+ "layer_types": [
45
+ "conv",
46
+ "conv",
47
+ "full_attention",
48
+ "conv",
49
+ "conv",
50
+ "full_attention",
51
+ "conv",
52
+ "conv",
53
+ "conv",
54
+ "full_attention",
55
+ "conv",
56
+ "conv",
57
+ "conv",
58
+ "full_attention",
59
+ "conv",
60
+ "conv",
61
+ "conv",
62
+ "full_attention",
63
+ "conv",
64
+ "conv",
65
+ "conv",
66
+ "full_attention",
67
+ "conv",
68
+ "conv",
69
+ "full_attention",
70
+ "conv",
71
+ "conv",
72
+ "full_attention",
73
+ "conv",
74
+ "conv"
75
+ ],
76
+ "max_position_embeddings": 128000,
77
+ "model_type": "lfm2",
78
+ "norm_eps": 1e-05,
79
+ "num_attention_heads": 32,
80
+ "num_heads": 32,
81
+ "num_hidden_layers": 30,
82
+ "num_key_value_heads": 8,
83
+ "rope_parameters": {
84
+ "rope_theta": 1000000.0,
85
+ "rope_type": "default"
86
+ },
87
+ "theta": 1000000.0,
88
+ "tie_embedding": true,
89
+ "use_cache": true,
90
+ "use_pos_enc": true,
91
+ "vocab_size": 65536
92
+ },
93
+ "tile_size": 512,
94
+ "transformers_version": "5.0.0.dev0",
95
+ "use_image_special_tokens": true,
96
+ "use_thumbnail": true,
97
+ "vision_config": {
98
+ "attention_dropout": 0.0,
99
+ "dtype": "bfloat16",
100
+ "hidden_act": "gelu_pytorch_tanh",
101
+ "hidden_size": 1152,
102
+ "intermediate_size": 4304,
103
+ "layer_norm_eps": 1e-06,
104
+ "model_type": "siglip2_vision_model",
105
+ "num_attention_heads": 16,
106
+ "num_channels": 3,
107
+ "num_hidden_layers": 27,
108
+ "num_patches": 256,
109
+ "patch_size": 16,
110
+ "vision_use_head": false
111
+ },
112
+ "transformers.js_config": {
113
+ "use_external_data_format": {
114
+ "vision_encoder.onnx": 1,
115
+ "decoder_model_merged.onnx": 6,
116
+ "embed_tokens.onnx": 1,
117
+ "vision_encoder_fp16.onnx": 1,
118
+ "decoder_model_merged_fp16.onnx": 3,
119
+ "embed_tokens_fp16.onnx": 1
120
+ },
121
+ "kv_cache_dtype": {
122
+ "q4f16": "float16",
123
+ "fp16": "float16"
124
+ }
125
+ }
126
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 7,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "5.0.0.dev0"
7
+ }
onnx/decoder_model_merged.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f31c4e3b9b1f1559becea58f2150d333502359cde2131c8c6c4bddb64072ff80
3
+ size 260567
onnx/decoder_model_merged.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c2f71e3977d0af0367f1d8712590675d409f6478a8d4908d113b034f36f27f0
3
+ size 2037882880
onnx/decoder_model_merged.onnx_data_1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b342d2168c8cf6567b112fffba58a050a6fc44d10dc92194b50a7d33e1f327c7
3
+ size 2051235840
onnx/decoder_model_merged.onnx_data_2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:341f3698484a47bfc6718b42cd445acafba9e6bb3dc575dfb9a141567898dc42
3
+ size 2026045440
onnx/decoder_model_merged.onnx_data_3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f4c07e637e85344e26dc0a4902b37d49abc6f8dc1f96021a10b16a83568231b
3
+ size 2093195264
onnx/decoder_model_merged.onnx_data_4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15265de0c5fcf8c2648ad521980d99db7c5d0cc999690050ead3f353f7e43f4d
3
+ size 1564622848
onnx/decoder_model_merged.onnx_data_5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d154a4f5bb57c10c2278110be6837c93ef38bd8a9a2f3a4ca337efc76120152e
3
+ size 536870912
onnx/decoder_model_merged_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d072748ad0cdbd07e2afa8533a042763f212ef0a6698ee8c5471eab92c6da71
3
+ size 260102
onnx/decoder_model_merged_fp16.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b326ecf5e6affc0570382211f049a9184f7d86e1feef4dc5f2c5408d8c02d5cd
3
+ size 2088599552
onnx/decoder_model_merged_fp16.onnx_data_1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b18540bfca2b70213a9828b4e5b37ad2d634ac4a644be04c88c50d3d0c6bd3ee
3
+ size 2059620352
onnx/decoder_model_merged_fp16.onnx_data_2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:88e73e2b2d4f27a958c732e59d8f8c8c29629cb44c8deda9edb9bb4e9d97fa02
3
+ size 1006706688
onnx/embed_tokens.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d96718279758de0ad236e2dd1d1e189e3aaf6371c993e1e6becd3072a5192b46
3
+ size 434
onnx/embed_tokens.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94f5772966a1069b0b87d040fb5f4103f336f919f4078de7dbee6f38c6d910a3
3
+ size 536870912
onnx/embed_tokens_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49b8ba54c33af88573e809a68d7c4aa63bd79d222802b4b5026929334715db13
3
+ size 637
onnx/embed_tokens_fp16.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e2b069f1789a900b7914409ac3626d5c67706e4eaf0e47717fba2571832b4f6
3
+ size 268435456
onnx/vision_encoder.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa40e8f423fd3d17585c92efb900d39b868b6aad0fb5d8d814473a962c6ff288
3
+ size 271987
onnx/vision_encoder.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cf6c90d27effacb87bab5985dbccdb7166d7663c499803dad4283ecdf271d3b
3
+ size 1718811584
onnx/vision_encoder_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f9f65db7a21cb176acaa89b756f85fbe158a96f79a25361dc316c38f6441be0
3
+ size 275260
onnx/vision_encoder_fp16.onnx_data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05b70383729c2dca02fd40108b099fb3fafc19130d956aca46ea556ad4001099
3
+ size 859405792
processor_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "crop_size": null,
4
+ "data_format": "channels_first",
5
+ "default_to_square": true,
6
+ "device": null,
7
+ "disable_grouping": null,
8
+ "do_center_crop": null,
9
+ "do_convert_rgb": null,
10
+ "do_image_splitting": true,
11
+ "do_normalize": true,
12
+ "do_pad": true,
13
+ "do_rescale": true,
14
+ "do_resize": false,
15
+ "downsample_factor": 2,
16
+ "encoder_patch_size": 16,
17
+ "image_mean": [
18
+ 0.5,
19
+ 0.5,
20
+ 0.5
21
+ ],
22
+ "image_processor_type": "Lfm2VlImageProcessorFast",
23
+ "image_seq_length": null,
24
+ "image_std": [
25
+ 0.5,
26
+ 0.5,
27
+ 0.5
28
+ ],
29
+ "input_data_format": null,
30
+ "max_image_tokens": 256,
31
+ "max_num_patches": 1024,
32
+ "max_pixels_tolerance": 2.0,
33
+ "max_tiles": 10,
34
+ "min_image_tokens": 64,
35
+ "min_tiles": 2,
36
+ "pad_size": null,
37
+ "processor_class": "Lfm2VlProcessor",
38
+ "resample": 3,
39
+ "rescale_factor": 0.00392156862745098,
40
+ "return_row_col_info": true,
41
+ "return_tensors": null,
42
+ "size": {
43
+ "height": 512,
44
+ "width": 512
45
+ },
46
+ "tile_size": 512,
47
+ "use_numpy_backend": true,
48
+ "use_thumbnail": true
49
+ },
50
+ "processor_class": "Lfm2VlProcessor"
51
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": null,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|startoftext|>",
5
+ "clean_up_tokenization_spaces": true,
6
+ "eos_token": "<|im_end|>",
7
+ "image_end_token": "<|image_end|>",
8
+ "image_start_token": "<|image_start|>",
9
+ "image_thumbnail": "<|img_thumbnail|>",
10
+ "image_token": "<image>",
11
+ "is_local": false,
12
+ "legacy": false,
13
+ "model_max_length": 1000000000000000019884624838656,
14
+ "model_specific_special_tokens": {
15
+ "image_end_token": "<|image_end|>",
16
+ "image_start_token": "<|image_start|>",
17
+ "image_thumbnail": "<|img_thumbnail|>",
18
+ "image_token": "<image>"
19
+ },
20
+ "pad_token": "<|pad|>",
21
+ "processor_class": "Lfm2VlProcessor",
22
+ "return_token_type_ids": false,
23
+ "sp_model_kwargs": {},
24
+ "spaces_between_special_tokens": false,
25
+ "tokenizer_class": "TokenizersBackend",
26
+ "use_default_system_prompt": false,
27
+ "use_fast": true,
28
+ "chat_template": "{{bos_token}}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n'}}{% if message['content'] is string %}{{ message['content'] }}{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' %}{{ '<image>' }}{% elif content['type'] == 'text' %}{{ content['text'] }}{% endif %}{% endfor %}{% endif %}{{'<|im_end|>\n'}}{% endfor %}{% if add_generation_prompt %}{{'<|im_start|>assistant\n' }}{% endif %}"
29
+ }