cyberunit/KHSLiveUSB
Updated
Feel Free
cyberunit
AI & ML interests
Autonomous Emergency Mobile AI
Recent Activity
updated
a model
8 days ago
cyberunit/KHSLiveUSB
published
a model
9 days ago
cyberunit/KHSLiveUSB
liked
a model
14 days ago
amd/Qwen2.5-Coder-7B-Instruct-onnx-ryzenai-npu
Organizations
None yet
updated
a
model
8 days ago
published
a
model
9 days ago
liked
a
model
14 days ago
upvoted
2
collections
18 days ago
reacted to
sagar007's
post with š„
19 days ago
Post
4137
š 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!
š§ 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
š 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)
š sagar007/multigemma
Benchmark Results:
- VQA Accuracy: 53.8%
- Works great for: animal detection, room identification, scene understanding
š **Try it yourself:**
- š¤ Model: sagar007/multigemma
- š® Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma
- š» 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! š
#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!
š§ 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
š 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)
š sagar007/multigemma
Benchmark Results:
- VQA Accuracy: 53.8%
- Works great for: animal detection, room identification, scene understanding
š **Try it yourself:**
- š¤ Model: sagar007/multigemma
- š® Demo: https://huggingface.co/spaces/sagar007/Multimodal-Gemma
- š» 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! š
#multimodal #gemma #clip #llava #vision-language #pytorch
- 9 replies
liked
a
model
23 days ago
liked
a
model
26 days ago
liked
a
Space
29 days ago
liked
a
model
about 1 month ago
updated
a
model
about 2 months ago
published
a
model
about 2 months ago
reacted to
danielhanchen's
post with š¤šā¤ļøš„
2 months ago
Post
8575
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.
š 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.
š Step-by-step Guide: https://docs.unsloth.ai/models/qwen3-next
Thinking GGUF: unsloth/Qwen3-Next-80B-A3B-Thinking-GGUF
liked
4
models
3 months ago
unsloth/Cosmos-Reason1-7B-GGUF
Image-Text-to-Text
ā¢
8B
ā¢
Updated
ā¢
1.16k
ā¢
12
NexaAI/Qwen2-Audio-7B-GGUF
Audio-Text-to-Text
ā¢
8B
ā¢
Updated
ā¢
6.66k
ā¢
169
unsloth/Qwen3-VL-4B-Thinking-GGUF
Image-Text-to-Text
ā¢
4B
ā¢
Updated
ā¢
9.45k
ā¢
19
ggml-org/Qwen3-VL-2B-Instruct-GGUF
2B
ā¢
Updated
ā¢
760
ā¢
8