Instructions to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic",
	filename="DeepSeek-R1-Distill-Llama-70B-heretic-Q3_K_M.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

Use Docker

docker model run hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

LM Studio
Jan
Ollama
How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with Ollama:
```
ollama run hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
```

Unsloth Studio new

How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic to start chatting

Docker Model Runner
How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with Docker Model Runner:
```
docker model run hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
```

Lemonade

How to use ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M

Run and chat with the model

lemonade run user.DeepSeek-R1-Distill-Llama-70B-heretic-Q4_K_M

List all available models

lemonade list

DeepSeek-R1-Distill-Llama-70B-heretic / Modelfile

ThalisAI

Add Ollama Modelfile with Llama 3.1 chat template fix

481e42a verified 3 months ago

raw

history blame contribute delete

1.46 kB

	# Modelfile for ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic
	#
	# This model is based on Llama 3.1 architecture and requires the Llama 3.1
	# chat template instead of the DeepSeek native template. The default template
	# from the GGUF metadata uses DeepSeek's fullwidth Unicode special tokens
	# which are not correctly handled by the Llama BPE tokenizer in some backends.
	#
	# Usage:
	# ollama create deepseek-r1-70b-heretic -f DeepSeek-R1-Distill-Llama-70B-heretic.Modelfile
	# ollama run deepseek-r1-70b-heretic "Hello!"
	#
	# To use a different quantization, change the FROM line:
	# FROM hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q4_K_M
	# FROM hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q3_K_M

	FROM hf.co/ThalisAI/DeepSeek-R1-Distill-Llama-70B-heretic:Q6_K

	TEMPLATE """{{- if .System }}<\|start_header_id\|>system<\|end_header_id\|>

	{{ .System }}<\|eot_id\|>{{ end }}
	{{- range $i, $_ := .Messages }}
	{{- $last := eq (len (slice $.Messages $i)) 1 }}
	{{- if eq .Role "user" }}<\|start_header_id\|>user<\|end_header_id\|>

	{{ .Content }}<\|eot_id\|>
	{{- else if eq .Role "assistant" }}<\|start_header_id\|>assistant<\|end_header_id\|>

	{{- if .Content }}{{ .Content }}
	{{- end }}{{- if not $last }}<\|eot_id\|>{{- end }}
	{{- end }}
	{{- if and $last (ne .Role "assistant") }}<\|start_header_id\|>assistant<\|end_header_id\|>

	{{ end }}
	{{- end }}"""

	PARAMETER stop <\|eot_id\|>
	PARAMETER stop <\|end_of_text\|>
	PARAMETER temperature 0.6
	PARAMETER top_p 0.95