Update requirements.txt

requirements.txt

@@ -1,8 +1,10 @@
-torch>=2.4.0
-torchvision>=0.19.0
+torch==2.6.0
+torchvision==0.21.0
+torchdata==0.10.1
+
 opencv-python>=4.9.0.80
-diffusers>=0.31.0
-transformers==4.51.3
+diffusers
+transformers
 #transformers==4.46.3 # was needed by llamallava used by i2v hunyuan before patch
 tokenizers>=0.20.3
 accelerate>=1.1.1
@@ -12,13 +14,16 @@ easydict
 ftfy
 dashscope
 imageio-ffmpeg
-# flash_attn    
-gradio==5.23.0   
+
+# flash_attn
+flash-attn @ https://github.com/Dao-AILab/flash-attention/releases/download/v2.7.4.post1/flash_attn-2.7.4.post1+cu12torch2.6cxx11abiFALSE-cp310-cp310-linux_x86_64.whl
+
+gradio==5.33.0   
 numpy>=1.23.5,<2
 einops
 moviepy==1.0.3
-mmgp==3.5.6
-peft==0.15.0
+mmgp==3.4.8
+peft==0.14.0
 mutagen
 pydantic==2.10.6
 decord
@@ -29,23 +34,8 @@ timm
 segment-anything
 omegaconf
 hydra-core
-librosa==0.11.0
+librosa
 loguru
 sentencepiece
 av
-opencv-python
-pygame>=2.1.0
-sounddevice>=0.4.0
-# rembg==2.0.65
-torchdiffeq >= 0.2.5
-tensordict >= 0.6.1
-open_clip_torch >= 2.29.0
-pyloudnorm
-misaki
-soundfile
-ffmpeg-python
-pyannote.audio
-pynvml
-huggingface_hub[hf_xet]
-# num2words
-# spacy
+# rembg==2.0.65