ostris/Flex.1-alpha
Text-to-Image
โข
Updated
โข
1.53k
โข
158
Martin Viewegger
Viewegger
AI & ML interests
None yet
Recent Activity
liked
a model
1 day ago
ostris/Flex.1-alpha
liked
a model
1 day ago
drewThomasson/fineTunedTTSModels
reacted
to
alibabasglab's
post
with ๐
4 days ago
Introducing open-sourced ClearerVoice-Studio. A powerful speech processing AI tool to dramatically improve your speech quality. Checkout demo page: https://huggingface.co/spaces/alibabasglab/ClearVoice and https://modelscope.cn/studios/iic/ClearerVoice-Studio. Give us a Star on Github: https://github.com/modelscope/ClearerVoice-Studio!
Organizations
None yet
Viewegger's activity
liked
2
models
1 day ago
reacted to
alibabasglab's
post with ๐
4 days ago
Post
1093
Introducing open-sourced ClearerVoice-Studio. A powerful speech processing AI tool to dramatically improve your speech quality. Checkout demo page:
alibabasglab/ClearVoice and https://modelscope.cn/studios/iic/ClearerVoice-Studio. Give us a Star on Github: https://github.com/modelscope/ClearerVoice-Studio!
reacted to
alibabasglab's
post with ๐
8 days ago
Post
5241
๐ ClearerVoice-Studio New Feature: Speech Super-Resolution with MossFormer2 ! ๐
Weโre excited to announce that ClearerVoice-Studio now supports speech super-resolution, powered by our latest MossFormer2-based model!
Whatโs New?
๐ Convert Low-Resolution to High-Resolution Audio:
Transform low-resolution audio (effective sampling rate โฅ 16 kHz) into crystal-clear, high-resolution audio at 48 kHz.
๐ค Cutting-Edge Technology:
Leverages the MossFormer2 model plus HiFi-GAN, optimised for generating high-quality audio with enhanced perceptual clarity.
๐ง Enhanced Listening Experience:
Perfect for speech enhancement, content restoration, and high-fidelity audio applications.
๐ Try It Out!
Upgrade to the latest version of ClearerVoice-Studio (https://github.com/modelscope/ClearerVoice-Studio) to experience this powerful feature. Check out the updated documentation and examples in our repository.
Let us know your thoughts, feedback, or feature requests in the Issues section.
Weโre excited to announce that ClearerVoice-Studio now supports speech super-resolution, powered by our latest MossFormer2-based model!
Whatโs New?
๐ Convert Low-Resolution to High-Resolution Audio:
Transform low-resolution audio (effective sampling rate โฅ 16 kHz) into crystal-clear, high-resolution audio at 48 kHz.
๐ค Cutting-Edge Technology:
Leverages the MossFormer2 model plus HiFi-GAN, optimised for generating high-quality audio with enhanced perceptual clarity.
๐ง Enhanced Listening Experience:
Perfect for speech enhancement, content restoration, and high-fidelity audio applications.
๐ Try It Out!
Upgrade to the latest version of ClearerVoice-Studio (https://github.com/modelscope/ClearerVoice-Studio) to experience this powerful feature. Check out the updated documentation and examples in our repository.
Let us know your thoughts, feedback, or feature requests in the Issues section.
liked
a
model
16 days ago
liked
2
models
about 1 month ago
Benefit using Lora + finetuned model
1
#1 opened about 1 month ago
by
Viewegger
liked
2
models
about 2 months ago
Model quality on cca 200hour of audio
2
#8 opened 2 months ago
by
Viewegger
Dataset size and output quality
#2 opened 2 months ago
by
Viewegger
Training process details
4
#2 opened 2 months ago
by
Nils11
liked
a
Space
2 months ago
reacted to
m-ric's
post with ๐ฅ
2 months ago
Post
789
๐๐ฟ๐ฒ ๐๐ฐ๐ฎ๐น๐ถ๐ป๐ด ๐น๐ฎ๐๐ ๐ผ๐๐ฒ๐ฟ? ๐ ๐ฟ๐ฒ๐ฝ๐ผ๐ฟ๐ ๐ณ๐ฟ๐ผ๐บ ๐๐ต๐ฒ ๐๐ป๐ณ๐ผ๐ฟ๐บ๐ฎ๐๐ถ๐ผ๐ป ๐ฎ๐ป๐ป๐ผ๐๐ป๐ฐ๐ฒ๐ฑ ๐๐ต๐ฎ๐ ๐ข๐ฝ๐ฒ๐ป๐๐ ๐ถ๐ ๐๐ฒ๐ฒ๐ถ๐ป๐ด ๐ฑ๐ถ๐บ๐ถ๐ป๐ถ๐๐ต๐ถ๐ป๐ด ๐ฟ๐ฒ๐๐๐ฟ๐ป๐ ๐ณ๐ฟ๐ผ๐บ ๐๐ฐ๐ฎ๐น๐ถ๐ป๐ด ๐๐ฝ ๐๐ต๐ฒ ๐ป๐ฒ๐
๐ ๐๐ฃ๐ง ๐บ๐ผ๐ฑ๐ฒ๐น๐.
๐ What are scaling laws? These are empiric laws that say "Every time you increase compute spent in training 10-fold, your LLM's performance will go up by a predictable tick". Of course, they apply only if you train your model with the right methods.
The image below illustrates it: they're from a paper by Google, "Scaling Autoregressive Models for Content-Rich Text-to-Image Generation", and they show how quality and instruction following of models improve when you scale the model up (which is equivalent to scaling up the compute spent in training).
โก๏ธ These scaling laws have immense impact: they triggered the largest gold rush ever, with companies pouring billions into scaling up theiur training. Microsoft and OpenAI spent 100B into their "Startgate" mega training cluster, due to start running in 2028.
๐ค So, what about these reports of scaling laws slowing down?
If they are true, they would mean a gigantic paradigm shift, as the hundreds of billions poured by AI companies into scaling could be a dead-end. โ๏ธ
But I doubt it: until the most recent publications, scaling laws showed no signs of weakness, and the researchers at the higher end of the scale-up seems to imply the scaling up continues.
Wait and see!
๐ What are scaling laws? These are empiric laws that say "Every time you increase compute spent in training 10-fold, your LLM's performance will go up by a predictable tick". Of course, they apply only if you train your model with the right methods.
The image below illustrates it: they're from a paper by Google, "Scaling Autoregressive Models for Content-Rich Text-to-Image Generation", and they show how quality and instruction following of models improve when you scale the model up (which is equivalent to scaling up the compute spent in training).
โก๏ธ These scaling laws have immense impact: they triggered the largest gold rush ever, with companies pouring billions into scaling up theiur training. Microsoft and OpenAI spent 100B into their "Startgate" mega training cluster, due to start running in 2028.
๐ค So, what about these reports of scaling laws slowing down?
If they are true, they would mean a gigantic paradigm shift, as the hundreds of billions poured by AI companies into scaling could be a dead-end. โ๏ธ
But I doubt it: until the most recent publications, scaling laws showed no signs of weakness, and the researchers at the higher end of the scale-up seems to imply the scaling up continues.
Wait and see!
- 1 reply
liked
a
model
2 months ago
upvoted
a
paper
3 months ago
reacted to
yongchanghao's
post with ๐ฅ
3 months ago
Post
3757
We just released a paper (NeuZip) that compresses VRAM in a lossless manner to run larger models. This should be particularly useful when VRAM is insufficient during training/inference. Specifically, we look inside each floating number and find that the exponents are highly compressible (as shown in the figure below).
Read more about the work at NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks (2410.20650)
Read more about the work at NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks (2410.20650)
liked
a
dataset
3 months ago
Size of the dataset?
8
#1 opened 3 months ago
by
Viewegger