NousResearch/Hermes-4.3-36B-GGUF
Text Generation
•
36B
•
Updated
•
5.12k
•
30
Thijs
ThijsL202
AI & ML interests
None yet
Recent Activity
liked
a model
7 days ago
NousResearch/Hermes-4.3-36B-GGUF
liked
a model
7 days ago
NousResearch/Hermes-4.3-36B
liked
a model
7 days ago
arcee-ai/Trinity-Mini
Organizations
None yet
liked
3
models
7 days ago
liked
a
model
14 days ago
liked
3
models
19 days ago
liked
a
model
about 1 month ago
liked
a
Space
2 months ago
liked
2
models
3 months ago
liked
a
Space
3 months ago
liked
3
models
3 months ago
reacted to
codelion's
post with 🔥
3 months ago
Post
6182
I recently worked on a LoRA that improves tool use in LLM. Thought the approach might interest folks here.
The issue I have had when trying to use some of the local LLMs with coding agents is this:
Me: "Find all API endpoints with authentication in this codebase"
LLM: "You should look for @app .route decorators and check if they have auth middleware..."
But I often want it to search the files and show me but the LLM doesn't trigger a tool use call.
To fine-tune it for tool use I combined two data sources:
1. Magpie scenarios - 5000+ diverse tasks (bug hunting, refactoring, security audits)
2. Real execution - Ran these on actual repos (FastAPI, Django, React) to get authentic tool responses
This ensures the model learns both breadth (many scenarios) and depth (real tool behavior).
Tools We Taught:
-
-
-
-
-
-
Improvements:
- Tool calling accuracy: 12% → 80%
- Correct parameters: 8% → 87%
- Multi-step tasks: 3% → 78%
- End-to-end completion: 5% → 80%
- Tools per task: 0.2 → 3.8
The LoRA really improves on intential tool call as an example consider the query: "Find ValueError in payment module"
The response proceeds as follows:
1. Calls
2. Gets 4 matches across 3 files
3. Calls
4. Analyzes context
5. Reports: "Found 3 ValueError instances: payment/processor.py:47 for invalid amount, payment/validator.py:23 for unsupported currency..."
Resources:
- Colab notebook https://colab.research.google.com/github/codelion/ellora/blob/main/Ellora_Recipe_3_Enhanced_Tool_Calling_and_Code_Understanding.ipynb
- Model - codelion/Llama-3.2-1B-Instruct-tool-calling-lora
- GitHub - https://github.com/codelion/ellora
The issue I have had when trying to use some of the local LLMs with coding agents is this:
Me: "Find all API endpoints with authentication in this codebase"
LLM: "You should look for @app .route decorators and check if they have auth middleware..."
But I often want it to search the files and show me but the LLM doesn't trigger a tool use call.
To fine-tune it for tool use I combined two data sources:
1. Magpie scenarios - 5000+ diverse tasks (bug hunting, refactoring, security audits)
2. Real execution - Ran these on actual repos (FastAPI, Django, React) to get authentic tool responses
This ensures the model learns both breadth (many scenarios) and depth (real tool behavior).
Tools We Taught:
-
read_file - Actually read file contents-
search_files - Regex/pattern search across codebases-
find_definition - Locate classes/functions-
analyze_imports - Dependency tracking-
list_directory - Explore structure-
run_tests - Execute test suitesImprovements:
- Tool calling accuracy: 12% → 80%
- Correct parameters: 8% → 87%
- Multi-step tasks: 3% → 78%
- End-to-end completion: 5% → 80%
- Tools per task: 0.2 → 3.8
The LoRA really improves on intential tool call as an example consider the query: "Find ValueError in payment module"
The response proceeds as follows:
1. Calls
search_files with pattern "ValueError"2. Gets 4 matches across 3 files
3. Calls
read_file on each match4. Analyzes context
5. Reports: "Found 3 ValueError instances: payment/processor.py:47 for invalid amount, payment/validator.py:23 for unsupported currency..."
Resources:
- Colab notebook https://colab.research.google.com/github/codelion/ellora/blob/main/Ellora_Recipe_3_Enhanced_Tool_Calling_and_Code_Understanding.ipynb
- Model - codelion/Llama-3.2-1B-Instruct-tool-calling-lora
- GitHub - https://github.com/codelion/ellora
liked
a
model
3 months ago
reacted to
dhruv3006's
post with 🔥🚀
3 months ago
Post
1406
Pair a vision grounding model with a reasoning LLM with Cua
Cua just shipped v0.4 of the Cua Agent framework with Composite Agents - you can now pair a vision/grounding model with a reasoning LLM using a simple modelA+modelB syntax. Best clicks + best plans.
The problem: every GUI model speaks a different dialect.
• some want pixel coordinates
• others want percentages
• a few spit out cursed tokens like <|loc095|>
We built a universal interface that works the same across Anthropic, OpenAI, Hugging Face, etc.:
agent = ComputerAgent(
model="anthropic/claude-3-5-sonnet-20241022",
tools=[computer]
)
But here’s the fun part: you can combine models by specialization.
Grounding model (sees + clicks) + Planning model (reasons + decides) →
agent = ComputerAgent(
model="huggingface-local/HelloKKMe/GTA1-7B+openai/gpt-4o",
tools=[computer]
)
This gives GUI skills to models that were never built for computer use. One handles the eyes/hands, the other the brain. Think driver + navigator working together.
Two specialists beat one generalist. We’ve got a ready-to-run notebook demo - curious what combos you all will try.
Github : https://github.com/trycua/cua
Blog : https://www.trycua.com/blog/composite-agents
Cua just shipped v0.4 of the Cua Agent framework with Composite Agents - you can now pair a vision/grounding model with a reasoning LLM using a simple modelA+modelB syntax. Best clicks + best plans.
The problem: every GUI model speaks a different dialect.
• some want pixel coordinates
• others want percentages
• a few spit out cursed tokens like <|loc095|>
We built a universal interface that works the same across Anthropic, OpenAI, Hugging Face, etc.:
agent = ComputerAgent(
model="anthropic/claude-3-5-sonnet-20241022",
tools=[computer]
)
But here’s the fun part: you can combine models by specialization.
Grounding model (sees + clicks) + Planning model (reasons + decides) →
agent = ComputerAgent(
model="huggingface-local/HelloKKMe/GTA1-7B+openai/gpt-4o",
tools=[computer]
)
This gives GUI skills to models that were never built for computer use. One handles the eyes/hands, the other the brain. Think driver + navigator working together.
Two specialists beat one generalist. We’ve got a ready-to-run notebook demo - curious what combos you all will try.
Github : https://github.com/trycua/cua
Blog : https://www.trycua.com/blog/composite-agents
liked
a
model
4 months ago
