Projects I've worked on (or contributed to)
mrfakename PRO
mrfakename
AI & ML interests
LLMs, TTS, & Open Source
Recent Activity
updated
a model
about 6 hours ago
mrfakename/versificator
updated
a model
about 6 hours ago
mrfakename/ReverseBERT-EmbeddingGemma-300M
published
a model
about 6 hours ago
mrfakename/ReverseBERT-EmbeddingGemma-300M
Organizations
The SAM Audio model licenses allow for redistribution so long as the original license files are included
The OpenF5 TTS model series (currently OpenF5 TTS Base - more variants coming soon 👀)
TTS models that support zero-shot voice cloning
-
MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis
Paper • 2502.18924 • Published • 16 -
MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer
Paper • 2409.00750 • Published • 5 -
F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching
Paper • 2410.06885 • Published • 46 -
StyleTTS-ZS: Efficient High-Quality Zero-Shot Text-to-Speech Synthesis with Distilled Time-Varying Style Diffusion
Paper • 2409.10058 • Published • 2
My spaces or spaces I worked featured on Spaces of the Week! Ones at the top are the oldest, newest at the bottom 🤗
With LAION
Ministral 3 models converted to the Llama format (without the vision encoder)
Models converted to the Llama format
-
mrfakename/Apriel-5B-Instruct-llamafied
Text Generation • 5B • Updated • 13 • 4 -
mrfakename/Apriel-5B-Base-llamafied
Text Generation • 5B • Updated • 13 -
llamafy/Qwen-Qwen2.5-1.5B-llamafied
Text Generation • 2B • Updated • 16 -
llamafy/Qwen-Qwen2.5-1.5B-Instruct-llamafied
Text Generation • 2B • Updated • 13
Experiments that didn't work out.
Projects I've worked on (or contributed to)
Ministral 3 models converted to the Llama format (without the vision encoder)
The SAM Audio model licenses allow for redistribution so long as the original license files are included
The OpenF5 TTS model series (currently OpenF5 TTS Base - more variants coming soon 👀)
TTS models that support zero-shot voice cloning
-
MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis
Paper • 2502.18924 • Published • 16 -
MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer
Paper • 2409.00750 • Published • 5 -
F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching
Paper • 2410.06885 • Published • 46 -
StyleTTS-ZS: Efficient High-Quality Zero-Shot Text-to-Speech Synthesis with Distilled Time-Varying Style Diffusion
Paper • 2409.10058 • Published • 2
Models converted to the Llama format
-
mrfakename/Apriel-5B-Instruct-llamafied
Text Generation • 5B • Updated • 13 • 4 -
mrfakename/Apriel-5B-Base-llamafied
Text Generation • 5B • Updated • 13 -
llamafy/Qwen-Qwen2.5-1.5B-llamafied
Text Generation • 2B • Updated • 16 -
llamafy/Qwen-Qwen2.5-1.5B-Instruct-llamafied
Text Generation • 2B • Updated • 13
My spaces or spaces I worked featured on Spaces of the Week! Ones at the top are the oldest, newest at the bottom 🤗
Experiments that didn't work out.
With LAION