PaddlePaddle/PP-OCRv6_medium_det_safetensors
Image-to-Text • 22M • Updated • 718 • 70
rondale sidbury
milezdeep13
AI & ML interests
https://ai.studio/apps/c7a3806a-be7e-4775-a77e-96892dc0277c
Recent Activity
new activity about 23 hours ago
unsloth/Kimi-K2.7-Code-GGUF:Cl/CL liked a model 1 day ago
PaddlePaddle/PP-OCRv6_medium_det_safetensors upvoted a collection 1 day ago
mahoney-context-protocolOrganizations
Cl/CL
1
#7 opened about 24 hours ago
by milezdeep13
liked a
model 1 day ago
upvoted a collection 1 day ago
upvoted a paper 1 day ago
liked a
model 1 day ago
upvoted a paper 10 days ago
google/gemma-3n-E4B-it-litert-preview
👍🧠 221
36
#2006 opened about 1 year ago
by 
celsowm
HuggingFaceTB/SmolLM2-135M
#10351 opened 10 days ago
by milezdeep13
published a
bucket 10 days ago
upvoted a collection 10 days ago
upvoted a paper 10 days ago
updated a
collection 10 days ago
published a
Space 10 days ago
EIP1967
#10333 opened 11 days ago
by milezdeep13
Qwen/Qwen2.5-Omni-7B
➕🔥 130
18
#1 opened about 1 year ago
by 
julien-c
Create vp9_superframe
#10331 opened 11 days ago
by milezdeep13
reacted to matteospanio's post with 🚀 11 days ago
Post
7041
🎶 Released mule-torch — an unofficial PyTorch port of MULE (SF-NFNet-F0), SiriusXM/Pandora's music-audio embedding model (McCallum et al., ISMIR 2022).
No retraining: I re-implemented the architecture in pure PyTorch and transferred the original TensorFlow weights, then checked it layer by layer against the genuine TF pipeline.
✅ End-to-end clip-embedding cosine 0.9999999 vs the original
✅ ONNX backbone parity < 1e-6
✅ 62.35M params (paper: ~62.4M)
✅ Batched, GPU-native, ONNX-exportable — none of which the original
🤗 Weights: matteospanio/mule
💻 Code: https://github.com/matteospanio/mule-torch
📦 PyPI: https://pypi.org/project/mule-torch/
The fun bug: parity was perfect through every conv but the block output was anti-correlated (cos = −1). Cause: the learnable skip-init gains couldn't be mapped by layer name (Keras scrambles the order) — they had to be recovered from the graph.
⚠️ Unofficial, community port — not affiliated with or endorsed by the original authors. All credit to them; please cite the paper. Weights inherit CC-BY-NC-4.0.
No retraining: I re-implemented the architecture in pure PyTorch and transferred the original TensorFlow weights, then checked it layer by layer against the genuine TF pipeline.
✅ End-to-end clip-embedding cosine 0.9999999 vs the original
✅ ONNX backbone parity < 1e-6
✅ 62.35M params (paper: ~62.4M)
✅ Batched, GPU-native, ONNX-exportable — none of which the original
Analysis pipeline doespip install mule-torchfrom mule_torch import MuleModel
emb = MuleModel.from_pretrained()(waveform) # (B, T)@16kHz -> (B, 1728)🤗 Weights: matteospanio/mule
💻 Code: https://github.com/matteospanio/mule-torch
📦 PyPI: https://pypi.org/project/mule-torch/
The fun bug: parity was perfect through every conv but the block output was anti-correlated (cos = −1). Cause: the learnable skip-init gains couldn't be mapped by layer name (Keras scrambles the order) — they had to be recovered from the graph.
⚠️ Unofficial, community port — not affiliated with or endorsed by the original authors. All credit to them; please cite the paper. Weights inherit CC-BY-NC-4.0.
kernels-community/activation
#10330 opened 11 days ago
by milezdeep13
liked a
model 11 days ago
upvoted a changelog 11 days ago
Hugging Face Changelog
Introducing Kernels
• 200