Video-Text-to-Text
Transformers
Safetensors
English
videochat_flash_qwen
image-feature-extraction
multimodal
custom_code
Eval Results (legacy)
Instructions to use OpenGVLab/VideoChat-Flash-Qwen2-7B_res448 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use OpenGVLab/VideoChat-Flash-Qwen2-7B_res448 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("OpenGVLab/VideoChat-Flash-Qwen2-7B_res448", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
Update mm_utils.py
Browse files- mm_utils.py +1 -5
mm_utils.py
CHANGED
|
@@ -378,11 +378,7 @@ def load_video(video_path, max_num_frames=512, media_dict=None): #, media_dict):
|
|
| 378 |
else:
|
| 379 |
clip = None
|
| 380 |
|
| 381 |
-
|
| 382 |
-
from petrel_client.client import Client
|
| 383 |
-
client = Client(conf_path='~/petreloss.conf')
|
| 384 |
-
else:
|
| 385 |
-
client = None
|
| 386 |
|
| 387 |
frames, frame_indices, fps, duration = VIDEO_READER_FUNCS[media_dict['video_read_type']](video_path=video_path, num_frames=max_num_frames, sample='dynamic_fps1', fix_start=None, min_num_frames=64, max_num_frames=max_num_frames, client=client, clip=clip, local_num_frames=8)
|
| 388 |
|
|
|
|
| 378 |
else:
|
| 379 |
clip = None
|
| 380 |
|
| 381 |
+
client = None
|
|
|
|
|
|
|
|
|
|
|
|
|
| 382 |
|
| 383 |
frames, frame_indices, fps, duration = VIDEO_READER_FUNCS[media_dict['video_read_type']](video_path=video_path, num_frames=max_num_frames, sample='dynamic_fps1', fix_start=None, min_num_frames=64, max_num_frames=max_num_frames, client=client, clip=clip, local_num_frames=8)
|
| 384 |
|