Tortoise17
Tortoise17
@xgdgsc conda install somoclu doesnt work. Its broken.
`Requested somoclu==1.7.5.1 from file:///home/user/.cache/pip/wheels/fc/32/42/e4e40e81e81fd51c04ed5aeb131500247c16ade3e108e14fd2/somoclu-1.7.5-py3-none-any.whl has inconsistent version: expected '1.7.5.1', but metadata has '1.7.5'`
@ this is working, and this 1.7.6 is stable without any error or missing function so far. with above command of conda, its correct.
@peterwittek after python 3.8, the version 1.7.5.1 is no more supportive, that is reason it fails and newer version is not working accordingly.
@MahmoudAshraf97 This is a great effort. I want to ask that in your .ipynb example, is it possible to export the word level timestamps file also for the transcript and...
@buscon does transfer learning from the input audio sample for human spoken to generate a specific human voice with model with giga_speech already works?
@buscon did you manage to find the way? or any success?
@wenhuach21 ``` from auto_round import AutoRound model_name_or_path = "Qwen3-Omni-30B-A3B-Instruct" ar = AutoRound( model=model_name_or_path, scheme="W4A16", iters=50, lr=5e-3, ) ``` output_dir = "./tmp_autoround" ar.quantize_and_save(output_dir, format="auto_round") This was command but, pr #863 has...
@wenhuach21 The next error I am facing is below. ``` 2025-09-29 18:36:27 INFO autoround.py L146: using MLLM mode for multimodal model. 2025-09-29 18:36:28 WARNING modeling_utils.py L4569: `torch_dtype` is deprecated! Use...