LLM EXPLORER 61,491 MODELS INDEXED

Qwen2 VL 7B Instruct by Qwen

By Qwen · 33 downloads
HF Repo ↗ Copy URL

Qwen2 VL 7B Instruct is an open-source language model by Qwen. Features: 7b LLM, VRAM: 16.7GB, Context: 32K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.22.

Base model:finetune:qwen/qwen2... Base model:qwen/qwen2-vl-7b-in...   Conversational   En   Instruct   Multimodal   Qwen2 vl   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct 

Qwen2 VL 7B Instruct Parameters and Internals

Model Type 
multimodal, text-generation
Use Cases 
Areas:
research, commercial applications
Limitations:
No audio support, Data timeliness issue post-June 2023, Limited capability in recognizing individuals/IPs, Weak in complex instructions, Object counting and spatial reasoning difficulties
Additional Notes 
Supports up to 20min video understanding, multilingual text understanding within images.
Supported Languages 
primaryLanguages (English, Chinese), additionalLanguages (Most European languages, Japanese, Korean, Arabic, Vietnamese), description (Multilingual support for text understanding in images.)
Training Details 
Data Volume:
Data updated until June 2023
Model Architecture:
Naive Dynamic Resolution and Multimodal Rotary Position Embedding (M-ROPE)
Input Output 
Input Format:
text, image, video
Accepted Modalities:
text, image, video
Output Format:
text
Performance Tips:
Set min/max pixels for optimal speed and memory usage
LLM NameQwen2 VL 7B Instruct
Repository 🤗https://huggingface.co/Qwen/Qwen2-VL-7B-Instruct 
Base Model(s)  Qwen2 VL 7B Instruct   Qwen/Qwen2-VL-7B-Instruct
Model Size7b
Required VRAM16.7 GB
Updated2024-08-30
MaintainerQwen
Model Typeqwen2_vl
Instruction-BasedYes
Model Files  3.9 GB: 1-of-5   3.9 GB: 2-of-5   3.9 GB: 3-of-5   3.9 GB: 4-of-5   1.1 GB: 5-of-5
Supported Languagesen
Model ArchitectureQwen2VLForConditionalGeneration
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.41.2
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size152064
Torch Data Typebfloat16
Errorsreplace

Quantized Models of the Qwen2 VL 7B Instruct

Model
Likes
Downloads
VRAM
...L 7B Instruct Unsloth Bnb 4bit1358606 GB
Qwen2 VL 7B Instruct AWQ22107526 GB
OlmOCR 7B 0225 Preview 4bit1225 GB
...L 7B Instruct Abliterated 8bit1298 GB
Qwen2 VL 7B Instruct Bnb 4bit69365 GB
Qwen2 VL 7B Instruct GPTQ Int41425887 GB
Qwen2 VL 7B Instruct GPTQ Int831126610 GB

Best Alternatives to Qwen2 VL 7B Instruct

Best Alternatives
Context / RAM
Downloads
Likes
OlmOCR 7B 0225 Preview32K / 16.6 GB18655706
OlmOCR 7B 0225 Preview FP832K / 10 GB1279
OlmOCR 7B 0225 Preview Bf1632K / 16.6 GB284
Qwen2 VL 7B Instruct32K / 16.6 GB9086
Qwen2 VL 7B Instruct Bf1632K / 16.6 GB504
...L 7B Instruct Unsloth Bnb 4bit32K / 6.8 GB586013
OlmOCR 7B 0225 Preview 4bit32K / 5.6 GB221
...L 7B Instruct Abliterated 8bit32K / 8.9 GB291
Qwen2 VL 7B Instruct Bnb 4bit32K / 5.9 GB9366
Qwen2 VL 7B Instruct 4bit32K / 4.7 GB6482
Note: green Score (e.g. "73.2") means that the model is better than Qwen/Qwen2-VL-7B-Instruct.