Instructions to use beyoru/EvolLLM-Linh with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use beyoru/EvolLLM-Linh with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="beyoru/EvolLLM-Linh")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("beyoru/EvolLLM-Linh")
model = AutoModelForCausalLM.from_pretrained("beyoru/EvolLLM-Linh", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use beyoru/EvolLLM-Linh with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "beyoru/EvolLLM-Linh"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "beyoru/EvolLLM-Linh",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/beyoru/EvolLLM-Linh

SGLang

How to use beyoru/EvolLLM-Linh with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "beyoru/EvolLLM-Linh" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "beyoru/EvolLLM-Linh",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "beyoru/EvolLLM-Linh" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "beyoru/EvolLLM-Linh",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use beyoru/EvolLLM-Linh with Docker Model Runner:
```
docker model run hf.co/beyoru/EvolLLM-Linh
```

beyoru commited on Oct 30, 2025

Commit

d4ed45e

verified ·

1 Parent(s): 68a89e8

Update README.md

Browse files

Files changed (1) hide show

README.md +26 -17

README.md CHANGED Viewed

@@ -1,28 +1,19 @@
----
-library_name: transformers
-tags:
-- tool
-- function-calling
-- agent
-- merge
-base_model:
-- Qwen/Qwen3-4B-Instruct-2507
-- beyoru/Qwen3-4B-I-1209
-- Qwen/Qwen3-4B-Thinking-2507
-datasets:
-- Salesforce/xlam-function-calling-60k
----
 library_name: transformers
 tags:
 - tool
 - function-calling
 - agent
 base_model:
 - Qwen/Qwen3-4B-Instruct-2507
 datasets:
 - Salesforce/xlam-function-calling-60k
 ---
 # 🧠 **Model Card — EvolLLM-Linh**
 ### **Model Overview**
@@ -42,11 +33,18 @@ It aims to enhance **robustness, accuracy, and dialogue coherence** of LLMs oper
 - Robust multi-turn dialogue consistency
 - Adaptive understanding of user preferences and intent shifts
 ---
 ### **Evaluation Comparison**
-| **Category**                    |  **EvolLLM-Linh** |  **GPT-OSS-20B**  | **xLAM-2-8b-fc-r** | **Qwen3-2507** |
 | ------------------------------- | :---------------: | :---------------: | :-------: | :-----------: |
 | SINGLE TURN – SINGLE FUNCTION   |       0.800       |       0.800       |    0.63   |      0.69     |
 | SINGLE TURN – PARALLEL FUNCTION |       0.660       |       0.620       |    0.16   |      0.51     |
@@ -60,12 +58,21 @@ It aims to enhance **robustness, accuracy, and dialogue coherence** of LLMs oper
 | ATOMIC TASK – LIST              |       0.920       |       0.900       |    0.84   |      0.78     |
 | ATOMIC TASK – OBJECT (DEEP)     |       0.580       |       0.520       |    0.32   |      0.36     |
 | ATOMIC TASK – OBJECT (SHORT)    |       0.800       |       0.960       |    0.70   |      0.56     |
-| **Overall Accuracy**            | **0.750**         |     **0.760**     |  **0.61** |    **0.64**   |
 ---
 ### **Leaderboard Reference**
-Both **EvolLLM-Linh** and **GPT-OSS-20B** are benchmarked using **[ACEBench](https://chenchen0103.github.io/ACEBench/)** — assessing **function calling**, **compositional reasoning**, and **multi-turn interaction**.
 Results are **internal benchmarks** aligned with ACEBench task categories.
 ---
@@ -83,6 +90,8 @@ Results are **internal benchmarks** aligned with ACEBench task categories.
   </a>
 </p>
 ### **License**
 **MIT License** — free for research and non-commercial use with attribution.
 © 2025 beyoru.

+---
 library_name: transformers
 tags:
 - tool
 - function-calling
 - agent
+- merge
 base_model:
 - Qwen/Qwen3-4B-Instruct-2507
+- beyoru/Qwen3-4B-I-1209
+- Qwen/Qwen3-4B-Thinking-2507
 datasets:
 - Salesforce/xlam-function-calling-60k
 ---
 # 🧠 **Model Card — EvolLLM-Linh**
 ### **Model Overview**
 - Robust multi-turn dialogue consistency
 - Adaptive understanding of user preferences and intent shifts
+<p align="center">
+  <img src="hyacine-hsr.gif" width="150">
+</p>
 ---
 ### **Evaluation Comparison**
+---
+| **Category**                    |  **EvolLLM-Linh** |  **GPT-OSS-20B**  | **Llama** | **Qwen-2507** |
 | ------------------------------- | :---------------: | :---------------: | :-------: | :-----------: |
 | SINGLE TURN – SINGLE FUNCTION   |       0.800       |       0.800       |    0.63   |      0.69     |
 | SINGLE TURN – PARALLEL FUNCTION |       0.660       |       0.620       |    0.16   |      0.51     |
 | ATOMIC TASK – LIST              |       0.920       |       0.900       |    0.84   |      0.78     |
 | ATOMIC TASK – OBJECT (DEEP)     |       0.580       |       0.520       |    0.32   |      0.36     |
 | ATOMIC TASK – OBJECT (SHORT)    |       0.800       |       0.960       |    0.70   |      0.56     |
+| **Overall Accuracy**            | **0.750 (75.0%)** | **0.760 (76.0%)** |  **0.61** |    **0.64**   |
+---
+> **Note:**
+> **We evaluate all models with the same configuration.**
+> If you find any incorrect or inconsistent result, please report it for verification.
+> This ensures transparency and reproducibility across benchmarks.
+<p align="center">
+  <img src="https://cdn-uploads.huggingface.co/production/uploads/65905af887944e494e37e09a/XB1XEInyfE3dyUNAGb5zF.webp" width="300">
+</p>
 ---
 ### **Leaderboard Reference**
+all model are benchmarked using **[ACEBench](https://chenchen0103.github.io/ACEBench/)** — assessing **function calling**, **compositional reasoning**, and **multi-turn interaction**.
 Results are **internal benchmarks** aligned with ACEBench task categories.
 ---
   </a>
 </p>
+## Notes:
+**We evaluate all models with a same configure**. IF there are incorrect result please report.
 ### **License**
 **MIT License** — free for research and non-commercial use with attribution.
 © 2025 beyoru.