nvidia/DeepSeek-V4-Flash-0731-NVFP4
Text Generation • 304B • Updated • 65 • 11
Jaiyam Sharma
dataplayer12
AI & ML interests
Computer Vision
Recent Activity
liked a model about 12 hours ago
nvidia/DeepSeek-V4-Flash-0731-NVFP4 liked a model about 1 month ago
moonshotai/Kimi-K3 liked a model about 2 months ago
nvidia/Nemotron-3-Embed-1B-NVFP4Organizations
liked a
model about 12 hours ago
liked a
model about 1 month ago
liked a
model about 2 months ago
upvoted an article about 2 months ago
Article
NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval
nvidia
• • 59 liked a
model about 2 months ago
liked a
model 8 months ago
liked a
model 10 months ago
`gemma3_text` is not supported in mlx
1
#1 opened 12 months ago
by dataplayer12
updated a
model over 1 year ago
published a
model over 1 year ago
updated a
model over 1 year ago
published a
model over 1 year ago
updated a
model over 1 year ago
published a
model over 1 year ago
liked a
model over 1 year ago
reacted to chansung's post with 👍 over 1 year ago
Post
2049
Simple summary on DeepSeek AI's Janus-Pro: A fresh take on multimodal AI!
It builds on its predecessor, Janus, by tweaking the training methodology rather than the model architecture. The result? Improved performance in understanding and generating multimodal data.
Janus-Pro uses a three-stage training strategy, similar to Janus, but with key modifications:
✦ Stage 1 & 2: Focus on separate training for specific objectives, rather than mixing data.
✦ Stage 3: Fine-tuning with a careful balance of multimodal data.
Benchmarks show Janus-Pro holds its own against specialized models like TokenFlow XL and MetaMorph, and other multimodal models like SD3 Medium and DALL-E 3.
The main limitation? Low image resolution (384x384). However, this seems like a strategic choice to focus on establishing a solid "recipe" for multimodal models. Future work will likely leverage this recipe and increased computing power to achieve higher resolutions.
It builds on its predecessor, Janus, by tweaking the training methodology rather than the model architecture. The result? Improved performance in understanding and generating multimodal data.
Janus-Pro uses a three-stage training strategy, similar to Janus, but with key modifications:
✦ Stage 1 & 2: Focus on separate training for specific objectives, rather than mixing data.
✦ Stage 3: Fine-tuning with a careful balance of multimodal data.
Benchmarks show Janus-Pro holds its own against specialized models like TokenFlow XL and MetaMorph, and other multimodal models like SD3 Medium and DALL-E 3.
The main limitation? Low image resolution (384x384). However, this seems like a strategic choice to focus on establishing a solid "recipe" for multimodal models. Future work will likely leverage this recipe and increased computing power to achieve higher resolutions.
updated a
model over 1 year ago
published a
model over 1 year ago
liked 2
models over 2 years ago
