Instructions to use ethicalabs/xLSTM-7b-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ethicalabs/xLSTM-7b-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="ethicalabs/xLSTM-7b-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ethicalabs/xLSTM-7b-Instruct")
model = AutoModelForCausalLM.from_pretrained("ethicalabs/xLSTM-7b-Instruct")
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 ethicalabs/xLSTM-7b-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ethicalabs/xLSTM-7b-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ethicalabs/xLSTM-7b-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ethicalabs/xLSTM-7b-Instruct

SGLang

How to use ethicalabs/xLSTM-7b-Instruct 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 "ethicalabs/xLSTM-7b-Instruct" \
    --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": "ethicalabs/xLSTM-7b-Instruct",
		"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 "ethicalabs/xLSTM-7b-Instruct" \
        --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": "ethicalabs/xLSTM-7b-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ethicalabs/xLSTM-7b-Instruct with Docker Model Runner:
```
docker model run hf.co/ethicalabs/xLSTM-7b-Instruct
```

mrs83 commited on Oct 15, 2025

Commit

30ecae2

verified ·

1 Parent(s): 0a1d72c

Update README.md

Browse files

Files changed (1) hide show

README.md +47 -7

README.md CHANGED Viewed

@@ -1,17 +1,18 @@
 ---
 base_model:
-- ethicalabs/xLSTM-7b-Instruct-PEFT
 library_name: transformers
 model_name: xlstm-7b-chatml
 tags:
-- base_model:adapter:NX-AI/xLSTM-7b
-- lora
 - sft
 - transformers
 - trl
 licence: license
 pipeline_tag: text-generation
 license: mit
 ---
 # Model Card for xLSTM-7b-Instruct
@@ -21,15 +22,56 @@ It has been trained using [TRL](https://github.com/huggingface/trl).
 ## Quick start
 ```python
-TODO
 ```
 ## Training procedure
 [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/ethicalabs-ai/xlstm-finetuning/runs/pfmf34a3)
 This model was trained with SFT.
 ### Framework versions
@@ -43,8 +85,6 @@ This model was trained with SFT.
 ## Citations
 Cite TRL as:
 ```bibtex

 ---
 base_model:
+- NX-AI/xLSTM-7b
 library_name: transformers
 model_name: xlstm-7b-chatml
 tags:
+- base_model:adapter:ethicalabs/xLSTM-7b-Instruct-PEFT
 - sft
 - transformers
 - trl
 licence: license
 pipeline_tag: text-generation
 license: mit
+datasets:
+- HuggingFaceH4/ultrachat_200k
 ---
 # Model Card for xLSTM-7b-Instruct
 ## Quick start
+For text generation you have to to pin specific pytorch versions [https://huggingface.co/datasets/John6666/forum1/blob/main/xlstm_1.md](https://huggingface.co/datasets/John6666/forum1/blob/main/xlstm_1.md)
+```shell
+pip install "torch==2.5.1" "torchvision==0.20.1" "torchaudio==2.5.1" --index-url https://download.pytorch.org/whl/cu124
+pip install "triton==3.4.0"  # >=3.1 is OK; 3.4.0 current as of Sep 2025
+pip install "mlstm-kernels==2.0.1" "xlstm==2.0.5"
+```
 ```python
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, AutoConfig
+MERGED_MODEL_PATH = "ethicalabs/xLSTM-7b-Instruct"
+# We apply a configuration that uses native, hardware-agnostic kernels.
+print("Defining a safe, native kernel configuration for compatibility...")
+safe_config = AutoConfig.from_pretrained(MERGED_MODEL_PATH, trust_remote_code=True)
+# Use the stable, native parallel kernel
+safe_config.chunkwise_kernel = "chunkwise--native_autograd"
+safe_config.sequence_kernel = "native_sequence__native"
+safe_config.step_kernel = "native"
+# This flag is still required for the HF implementation to avoid unpacking errors
+safe_config.return_last_states = False
+# Load the final, merged model with the safe config (no quantization)
+print("Loading the final, merged model in bfloat16 (no quantization for compatibility)...")
+final_model = AutoModelForCausalLM.from_pretrained(
+    MERGED_MODEL_PATH,
+    device_map="auto",
+    torch_dtype=torch.bfloat16,
+    trust_remote_code=True,
+    config=safe_config
+)
+final_tokenizer = AutoTokenizer.from_pretrained(MERGED_MODEL_PATH)
+# The tokenizer needs to know which token to use for padding.
+if final_tokenizer.pad_token is None:
+    final_tokenizer.pad_token = final_tokenizer.eos_token
+print("Padding token has been set.")
+# Set the model to evaluation mode
+final_model.eval()
 ```
 ## Training procedure
 [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/ethicalabs-ai/xlstm-finetuning/runs/pfmf34a3)
 This model was trained with SFT.
 ### Framework versions
 ## Citations
 Cite TRL as:
 ```bibtex