[GH-ISSUE #65] How to output Chinese using deepsek-vl-1.3b-chat #49

New issue

Closed

opened 2026-10-26 21:10:51 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-26 21:10:51 +00:00

Owner

Originally created by @Kenneth-X on GitHub (Feb 7, 2025).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-VL2/issues/65

How to output Chinese using deepsek-vl-1.3b-chat ? Here is my code and result

from lmdeploy import pipeline
from lmdeploy.vl import load_image
from lmdeploy.vl.constants import IMAGE_TOKEN

pipe = pipeline('deepseek-ai/deepseek-vl-1.3b-chat')

image = load_image('test.jpg')
response = pipe((f'用中文描述这个图片{IMAGE_TOKEN}', image))
print(response.text)

[WARNING] gemm_config.in is not found; using default GEMM algo                                                                                                           
2025-02-07 17:53:48,894 - lmdeploy - WARNING - async_engine.py:505 - GenerationConfig: GenerationConfig(n=1, max_new_tokens=512, do_sample=False, top_p=1.0, top_k=50, min_p=0.0, temperature=0.8, repetition_penalty=1.0, ignore_eos=False, random_seed=None, stop_words=None, bad_words=None, stop_token_ids=None, bad_token_ids=None, min_new_tokens=None, skip_special_tokens=True, logprobs=None, response_format=None, logits_processors=None)
2025-02-07 17:53:48,894 - lmdeploy - WARNING - async_engine.py:506 - Since v0.6.0, lmdeploy add `do_sample` in GenerationConfig. It defaults to False, meaning greedy decoding. Please set `do_sample=True` if sampling  decoding is needed

**The image is a photograph of a street scene. On the right side of the street, there is a row of trees with green leaves, and a sidewalk runs parallel to the road. On the left side, there are cars driving on the road, and a bus is visible on the right side of the street. The road is marked with white lines, and there are traffic lights at the intersection. The sky is clear, and the overall atmosphere of the image is bright and sunny. There are no texts or discernible brands in the image.**

Originally created by @Kenneth-X on GitHub (Feb 7, 2025). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-VL2/issues/65 How to output Chinese using deepsek-vl-1.3b-chat ? Here is my code and result ```python from lmdeploy import pipeline from lmdeploy.vl import load_image from lmdeploy.vl.constants import IMAGE_TOKEN pipe = pipeline('deepseek-ai/deepseek-vl-1.3b-chat') image = load_image('test.jpg') response = pipe((f'用中文描述这个图片{IMAGE_TOKEN}', image)) print(response.text) ``` ```shell [WARNING] gemm_config.in is not found; using default GEMM algo 2025-02-07 17:53:48,894 - lmdeploy - WARNING - async_engine.py:505 - GenerationConfig: GenerationConfig(n=1, max_new_tokens=512, do_sample=False, top_p=1.0, top_k=50, min_p=0.0, temperature=0.8, repetition_penalty=1.0, ignore_eos=False, random_seed=None, stop_words=None, bad_words=None, stop_token_ids=None, bad_token_ids=None, min_new_tokens=None, skip_special_tokens=True, logprobs=None, response_format=None, logits_processors=None) 2025-02-07 17:53:48,894 - lmdeploy - WARNING - async_engine.py:506 - Since v0.6.0, lmdeploy add `do_sample` in GenerationConfig. It defaults to False, meaning greedy decoding. Please set `do_sample=True` if sampling decoding is needed **The image is a photograph of a street scene. On the right side of the street, there is a row of trees with green leaves, and a sidewalk runs parallel to the road. On the left side, there are cars driving on the road, and a bus is visible on the right side of the street. The road is marked with white lines, and there are traffic lights at the intersection. The sky is clear, and the overall atmosphere of the image is bright and sunny. There are no texts or discernible brands in the image.** ```

bsakaguchi closed this issue

2026-10-26 21:10:52 +00:00

No labels

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

deepseek-ai/DeepSeek-VL2#49

No description provided.

Rows
Columns