InternVL2-2B

Original model repository: OpenGVLab/InternVL2-2B

Model Introduction

InternVL2-2B is an instruction-tuned Vision-Language Model (VLM) for image understanding and text generation. It combines an InternViT-300M vision encoder, an MLP projector, and InternLM2-Chat-1.8B as its language model. Typical applications include visual question answering, OCR, image description, document and chart understanding, and multimodal dialogue.

Deployment Metrics

Model Parameters

Metric	Value
Total model parameters	2.206B
Vision model (ViT) parameters	316.6M
Language model (LM) parameters	1.889B

Parameter counts are calculated from the tensors stored in the upstream checkpoint.

Performance Metrics

Test Configuration

Metric	Value
Platform	Matrix6P
Data type	W8A8
ViT image size	448 × 448
Sequence length	512
Maximum context length	1024
BPU cores (ViT / Prefill / Decode)	4 / 4 / 4

Performance Results

Metric	Value
ViT latency	41.535 ms
Time to first token (TTFT)	84.448 ms
Prefill throughput	13,265.273 tokens/s
Decode throughput	70.473 tokens/s

Memory Usage

Metric	Value
BPU memory	2.4 GB
CPU memory	0.79 GB

Note: TTFT includes preprocessing and ViT latency. Memory values represent the peak memory usage measured during the specified performance test.