amd/Qwen2.5-Coder-7B-Instruct-onnx-ryzenai-npu
Text Generation
โข
Updated
โข
28
โข
1
Feel Free
cyberunit
AI & ML interests
Autonomous Emergency Mobile AI
Recent Activity
liked
a model
3 days ago
amd/Qwen2.5-Coder-7B-Instruct-onnx-ryzenai-npu
upvoted
a
collection
7 days ago
Ryzen AI 1.6 Whisper NPU Optimized ONNX models
upvoted
a
collection
7 days ago
Ryzen-AI-1.6-NPU-LLM
Organizations
None yet
liked
a
model
3 days ago
upvoted
2
collections
7 days ago
reacted to
sagar007's
post with ๐ฅ
9 days ago
Post
4116
๐ I built a Multimodal Vision-Language Model from using Gemma-270M + CLIP!
Just finished training my multimodal model on the full LLaVA-Instruct-150K dataset (157K samples) and wanted to share the results!
๐ง What I Built:
A vision-language model that can understand images and answer questions about them, combining:
- Google Gemma-3-270M (language)
- OpenAI CLIP ViT-Large/14 (vision)
- LoRA fine-tuning for efficiency
๐ Training Stats:
- 157,712 training samples (full LLaVA dataset)
- 3 epochs on A100 40GB
- ~9 hours training time
- Final loss: 1.333 training / 1.430 validation
- Only 18.6M trainable params (3.4% of 539M total)
๐ sagar007/multigemma
Benchmark Results:
- VQA Accuracy: 53.8%
- Works great for: animal detection, room identification, scene understanding
๐ **Try it yourself:**
- ๐ค Model: sagar007/multigemma
- ๐ฎ Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma
- ๐ป GitHub: https://github.com/sagar431/multimodal-gemma-270m
Built with PyTorch Lightning + MLflow for experiment tracking. Full MLOps pipeline with CI/CD!
Would love to hear your feedback! ๐
#multimodal #gemma #clip #llava #vision-language #pytorch
Just finished training my multimodal model on the full LLaVA-Instruct-150K dataset (157K samples) and wanted to share the results!
๐ง What I Built:
A vision-language model that can understand images and answer questions about them, combining:
- Google Gemma-3-270M (language)
- OpenAI CLIP ViT-Large/14 (vision)
- LoRA fine-tuning for efficiency
๐ Training Stats:
- 157,712 training samples (full LLaVA dataset)
- 3 epochs on A100 40GB
- ~9 hours training time
- Final loss: 1.333 training / 1.430 validation
- Only 18.6M trainable params (3.4% of 539M total)
๐ sagar007/multigemma
Benchmark Results:
- VQA Accuracy: 53.8%
- Works great for: animal detection, room identification, scene understanding
๐ **Try it yourself:**
- ๐ค Model: sagar007/multigemma
- ๐ฎ Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma
- ๐ป GitHub: https://github.com/sagar431/multimodal-gemma-270m
Built with PyTorch Lightning + MLflow for experiment tracking. Full MLOps pipeline with CI/CD!
Would love to hear your feedback! ๐
#multimodal #gemma #clip #llava #vision-language #pytorch
- 9 replies
liked
a
model
12 days ago
liked
a
model
15 days ago
liked
a
Space
18 days ago
liked
a
model
21 days ago
updated
a
model
about 2 months ago
published
a
model
about 2 months ago
reacted to
danielhanchen's
post with ๐ค๐โค๏ธ๐ฅ
about 2 months ago
Post
8570
Qwen3-Next can now be Run locally! (30GB RAM)
Instruct GGUF: unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF
The models come in Thinking and Instruct versions and utilize a new architecture, allowing it to have ~10x faster inference than Qwen32B.
๐ Step-by-step Guide: https://docs.unsloth.ai/models/qwen3-next
Thinking GGUF: unsloth/Qwen3-Next-80B-A3B-Thinking-GGUF
Instruct GGUF: unsloth/Qwen3-Next-80B-A3B-Instruct-GGUF
The models come in Thinking and Instruct versions and utilize a new architecture, allowing it to have ~10x faster inference than Qwen32B.
๐ Step-by-step Guide: https://docs.unsloth.ai/models/qwen3-next
Thinking GGUF: unsloth/Qwen3-Next-80B-A3B-Thinking-GGUF
liked
5
models
3 months ago
unsloth/Cosmos-Reason1-7B-GGUF
Image-Text-to-Text
โข
8B
โข
Updated
โข
1.14k
โข
12
NexaAI/Qwen2-Audio-7B-GGUF
Audio-Text-to-Text
โข
8B
โข
Updated
โข
6.12k
โข
168
unsloth/Qwen3-VL-4B-Thinking-GGUF
Image-Text-to-Text
โข
4B
โข
Updated
โข
5.26k
โข
18
ggml-org/Qwen3-VL-2B-Instruct-GGUF
2B
โข
Updated
โข
666
โข
7
ggml-org/LightOnOCR-1B-1025-GGUF
0.8B
โข
Updated
โข
848
โข
10
reacted to
merve's
post with ๐
3 months ago
Post
8315
deepseek-ai/DeepSeek-OCR is out! ๐ฅ my take โคต๏ธ
> pretty insane it can parse and re-render charts in HTML
> it uses CLIP and SAM features concatenated, so better grounding
> very efficient per vision tokens/performance ratio
> covers 100 languages
> pretty insane it can parse and re-render charts in HTML
> it uses CLIP and SAM features concatenated, so better grounding
> very efficient per vision tokens/performance ratio
> covers 100 languages
- 4 replies