Visual Question Answering
Transformers
Safetensors
English
videollama2_mistral
text-generation
multimodal large language model
large video-language model
Instructions to use ccclemenfff/AVL with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ccclemenfff/AVL with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("visual-question-answering", model="ccclemenfff/AVL")# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("ccclemenfff/AVL", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Commit ·
723daaf
1
Parent(s): e65359a
fix: disable flash attention 2 for CLIPVisionModel
Browse files- handler.py +0 -1
- videollama2/model/encoder.py +1 -1
handler.py
CHANGED
|
@@ -3,7 +3,6 @@ import base64
|
|
| 3 |
import tempfile
|
| 4 |
import os
|
| 5 |
import sys
|
| 6 |
-
os.environ["FLASH_ATTENTION_2_ENABLED"] = "false"
|
| 7 |
|
| 8 |
|
| 9 |
# 确保能导入 videollama2 模块(模型代码需要放同目录或已安装)
|
|
|
|
| 3 |
import tempfile
|
| 4 |
import os
|
| 5 |
import sys
|
|
|
|
| 6 |
|
| 7 |
|
| 8 |
# 确保能导入 videollama2 模块(模型代码需要放同目录或已安装)
|
videollama2/model/encoder.py
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
import os
|
| 2 |
-
|
| 3 |
import torch
|
| 4 |
import torch.nn as nn
|
| 5 |
|
|
|
|
| 1 |
import os
|
| 2 |
+
os.environ["TRANSFORMERS_NO_FLASH_ATTN_2"] = "1"
|
| 3 |
import torch
|
| 4 |
import torch.nn as nn
|
| 5 |
|