liuzhaochen02
commited on
Commit
•
8375f21
1
Parent(s):
673cd54
initial commit
Browse files- README.md +21 -0
- config.json +41 -0
- generation_config.json +7 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +24 -0
- tokenizer.model +3 -0
- tokenizer_config.json +37 -0
README.md
CHANGED
@@ -1,3 +1,24 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
+
datasets:
|
4 |
+
- 558K filtered image-text pairs from LAION/CC/SBU, captioned by BLIP.
|
5 |
+
- 158K GPT-generated multimodal instruction-following data.
|
6 |
+
- 450K academic-task-oriented VQA data mixture.
|
7 |
+
- 40K ShareGPT data.
|
8 |
+
tags:
|
9 |
+
- MobileVLM
|
10 |
---
|
11 |
+
## Model Summery
|
12 |
+
MobileVLM is a competent multimodal vision language model (MMVLM) targeted to run on mobile devices. It is an amalgamation of a myriad of architectural designs and techniques that are mobile-oriented, which comprises a set of language models at the scale of 1.4B and 2.7B parameters, trained from scratch, a multimodal vision model that is pre-trained in the CLIP fashion, cross-modality interaction via an efficient projector. We evaluate MobileVLM on several typical VLM benchmarks. Our models demonstrate on par performance compared with a few much larger models. More importantly, we measure the inference speed on both a Qualcomm Snapdragon 888 CPU and an NVIDIA Jeston Orin GPU, and we obtain state-of-the-art performance of 21.5 tokens and 65.3 tokens per second, respectively.
|
13 |
+
|
14 |
+
The MobileVLM-1.7B was built on our [MobileLLaMA-1.4B-Chat](](https://huggingface.co/mtgv/MobileLLaMA-1.4B-Chat)) to facilitate the off-the-shelf deployment.
|
15 |
+
|
16 |
+
## Model Sources
|
17 |
+
- Repository: https://github.com/Meituan-AutoML/MobileVLM
|
18 |
+
- Paper: https://arxiv.org/abs/2312.16886
|
19 |
+
|
20 |
+
## How to Get Started with the Model
|
21 |
+
Inference examples can be found at [Github](https://github.com/Meituan-AutoML/MobileVLM).
|
22 |
+
|
23 |
+
## Training Details
|
24 |
+
Please refer to our paper: [MobileVLM: A Fast, Strong and Open Vision Language Assistant for Mobile Devices](https://arxiv.org/pdf/2312.16886.pdf)
|
config.json
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "mtgv/MobileVLM-1.7B",
|
3 |
+
"architectures": [
|
4 |
+
"MobileLlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"bos_token_id": 1,
|
7 |
+
"eos_token_id": 2,
|
8 |
+
"freeze_mm_mlp_adapter": false,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 2048,
|
11 |
+
"image_aspect_ratio": "pad",
|
12 |
+
"image_grid_pinpoints": null,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 5632,
|
15 |
+
"max_position_embeddings": 2048,
|
16 |
+
"max_sequence_length": 2048,
|
17 |
+
"mm_hidden_size": 1024,
|
18 |
+
"mm_projector_type": "ldpnet",
|
19 |
+
"mm_use_im_patch_token": false,
|
20 |
+
"mm_use_im_start_end": false,
|
21 |
+
"mm_vision_select_feature": "patch",
|
22 |
+
"mm_vision_select_layer": -2,
|
23 |
+
"mm_vision_tower": "openai/clip-vit-large-patch14-336",
|
24 |
+
"model_type": "mobilevlm",
|
25 |
+
"num_attention_heads": 16,
|
26 |
+
"num_hidden_layers": 24,
|
27 |
+
"num_key_value_heads": 16,
|
28 |
+
"pad_token_id": 0,
|
29 |
+
"pretraining_tp": 1,
|
30 |
+
"rms_norm_eps": 1e-06,
|
31 |
+
"rope_scaling": null,
|
32 |
+
"rope_theta": 10000.0,
|
33 |
+
"tie_word_embeddings": false,
|
34 |
+
"torch_dtype": "bfloat16",
|
35 |
+
"transformers_version": "4.33.1",
|
36 |
+
"tune_mm_mlp_adapter": false,
|
37 |
+
"use_cache": true,
|
38 |
+
"use_mm_proj": true,
|
39 |
+
"vision_tower_type": "clip",
|
40 |
+
"vocab_size": 32000
|
41 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ae73147f336eb9ecade7300d75642b0ea00a92ae23d8d30466a9d5a3aef2d592
|
3 |
+
size 3373705013
|
special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": true,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "<unk>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": true,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
3 |
+
size 499723
|
tokenizer_config.json
ADDED
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": true,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"bos_token": {
|
5 |
+
"__type": "AddedToken",
|
6 |
+
"content": "<s>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"clean_up_tokenization_spaces": false,
|
13 |
+
"eos_token": {
|
14 |
+
"__type": "AddedToken",
|
15 |
+
"content": "</s>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": true,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false
|
20 |
+
},
|
21 |
+
"legacy": true,
|
22 |
+
"model_max_length": 2048,
|
23 |
+
"pad_token": null,
|
24 |
+
"padding_side": "right",
|
25 |
+
"sp_model_kwargs": {},
|
26 |
+
"spaces_between_special_tokens": false,
|
27 |
+
"tokenizer_class": "LlamaTokenizer",
|
28 |
+
"unk_token": {
|
29 |
+
"__type": "AddedToken",
|
30 |
+
"content": "<unk>",
|
31 |
+
"lstrip": false,
|
32 |
+
"normalized": true,
|
33 |
+
"rstrip": false,
|
34 |
+
"single_word": false
|
35 |
+
},
|
36 |
+
"use_default_system_prompt": true
|
37 |
+
}
|