LiquidAI/ifstruct-v1.0
Viewer âĸ Updated âĸ 2k âĸ 33 âĸ 24
đ¤ Open to Collab
LH-Tech AI
LH-Tech-AI
AI & ML interests
Small AI and ML models. Trained by myself. Completely OpenSource. For you. | Reddit: https://www.reddit.com/user/LH-Tech_AI/
Recent Activity
liked a dataset about 16 hours ago
LiquidAI/ifstruct-v1.0 liked a dataset about 16 hours ago
Rapidata/svg-benchmark liked a dataset about 17 hours ago
glaiveai/glaive-function-calling-v2Organizations
liked 2
datasets about 16 hours ago
liked a
dataset about 17 hours ago
liked a
dataset about 18 hours ago
liked a
Space 1 day ago
liked a
dataset 1 day ago
liked a
model 1 day ago
liked 4
datasets 2 days ago
liked a
model 2 days ago
upvoted a collection 3 days ago
liked a
dataset 3 days ago
upvoted 2
collections 3 days ago
replied to Banaxi-Tech's post 3 days ago
Hmmm...this is a bit over half. Just keep it running and see in the end. Good luck.
Show me benchmarks and your exact config maybe đ¤
reacted to AxionLab-official's post with â¤ī¸ 3 days ago
Post
4897
â ī¸ Community Notice
We would like to clarify that SupraLabs has no affiliation, partnership, or connection whatsoever with "SupraLarps" or its members.
Please avoid interacting with their organization, repositories, or Spaces under the assumption that they are associated with us.
We are currently aware of the situation and have already contacted the appropriate channels to address it.
Thank you to everyone who continues to support SupraLabs. â¤ī¸
We would like to clarify that SupraLabs has no affiliation, partnership, or connection whatsoever with "SupraLarps" or its members.
Please avoid interacting with their organization, repositories, or Spaces under the assumption that they are associated with us.
We are currently aware of the situation and have already contacted the appropriate channels to address it.
Thank you to everyone who continues to support SupraLabs. â¤ī¸
- 11 replies
Report
đ 4
8
#1 opened 4 days ago
by 
finnianx
reacted to pankajpandey-dev's post with đ 4 days ago
Post
7770
đŽđŗ New in my Hindi LLM Series: Gemma-4 E4B, fine-tuned for Hindi â and it runs on your laptop's CPU.
I fine-tuned Google's new Gemma-4 E4B on ~10k Hindi instruction pairs (AI4Bharat: anudesh + dolly) using Unsloth + LoRA, on a single L4 GPU.
Then I ran an honest side-by-side eval: base Gemma-4 vs my fine-tune, across 25 Hindi prompts. The results were interesting đ
â My fine-tune is more concise â ask for "3 tips" and it gives exactly 3. Base writes a 1,200-character essay.
â Pure native Hindi â base keeps slipping into English ("ā¤¸ā¤ā¤¤āĨā¤˛ā¤ŋā¤¤ ā¤ā¤šā¤žā¤° (Eat a Balanced Diet)", "ā¤¤ā¤žā¤°ā¤ž (Star)"). My fine-tune stays in clean Hindi.
â Tighter instruction-following â ask for a "short message" and it gives one, not a menu of options.
âī¸ And to be honest: base Gemma-4 is more detailed and comprehensive. I didn't build a "smarter" model â I built a focused, Hindi-native, edge-friendly one that runs as a 5GB GGUF (Q4) on CPU.
đ Try it:
Live demo (CPU): pankajpandey-dev/gemma-4-e4b-hindi-demo
GGUF (Ollama/llama.cpp): pankajpandey-dev/gemma-4-e4b-hindi-instruct-GGUF
16-bit model: pankajpandey-dev/gemma-4-e4b-hindi-instruct
Built with @unsloth ¡ Data by @ai4bharat đ
#Hindi #LLM #Gemma #Unsloth #IndicNLP #GGUF
I fine-tuned Google's new Gemma-4 E4B on ~10k Hindi instruction pairs (AI4Bharat: anudesh + dolly) using Unsloth + LoRA, on a single L4 GPU.
Then I ran an honest side-by-side eval: base Gemma-4 vs my fine-tune, across 25 Hindi prompts. The results were interesting đ
â My fine-tune is more concise â ask for "3 tips" and it gives exactly 3. Base writes a 1,200-character essay.
â Pure native Hindi â base keeps slipping into English ("ā¤¸ā¤ā¤¤āĨā¤˛ā¤ŋā¤¤ ā¤ā¤šā¤žā¤° (Eat a Balanced Diet)", "ā¤¤ā¤žā¤°ā¤ž (Star)"). My fine-tune stays in clean Hindi.
â Tighter instruction-following â ask for a "short message" and it gives one, not a menu of options.
âī¸ And to be honest: base Gemma-4 is more detailed and comprehensive. I didn't build a "smarter" model â I built a focused, Hindi-native, edge-friendly one that runs as a 5GB GGUF (Q4) on CPU.
đ Try it:
Live demo (CPU): pankajpandey-dev/gemma-4-e4b-hindi-demo
GGUF (Ollama/llama.cpp): pankajpandey-dev/gemma-4-e4b-hindi-instruct-GGUF
16-bit model: pankajpandey-dev/gemma-4-e4b-hindi-instruct
Built with @unsloth ¡ Data by @ai4bharat đ
#Hindi #LLM #Gemma #Unsloth #IndicNLP #GGUF
- 12 replies