Instructions to use ProfEngel/OwlLM2-e2b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- llama-cpp-python
How to use ProfEngel/OwlLM2-e2b with llama-cpp-python:
# !pip install llama-cpp-python from llama_cpp import Llama llm = Llama.from_pretrained( repo_id="ProfEngel/OwlLM2-e2b", filename="OwlLM2-e2b.Q8_0.gguf", )
llm.create_chat_completion( messages = "\"The answer to the universe is 42\"" )
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- llama.cpp
How to use ProfEngel/OwlLM2-e2b with llama.cpp:
Install from brew
brew install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf ProfEngel/OwlLM2-e2b:Q8_0 # Run inference directly in the terminal: llama-cli -hf ProfEngel/OwlLM2-e2b:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama-server -hf ProfEngel/OwlLM2-e2b:Q8_0 # Run inference directly in the terminal: llama-cli -hf ProfEngel/OwlLM2-e2b:Q8_0
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 ProfEngel/OwlLM2-e2b:Q8_0 # Run inference directly in the terminal: ./llama-cli -hf ProfEngel/OwlLM2-e2b:Q8_0
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 ProfEngel/OwlLM2-e2b:Q8_0 # Run inference directly in the terminal: ./build/bin/llama-cli -hf ProfEngel/OwlLM2-e2b:Q8_0
Use Docker
docker model run hf.co/ProfEngel/OwlLM2-e2b:Q8_0
- LM Studio
- Jan
- Ollama
How to use ProfEngel/OwlLM2-e2b with Ollama:
ollama run hf.co/ProfEngel/OwlLM2-e2b:Q8_0
- Unsloth Studio new
How to use ProfEngel/OwlLM2-e2b 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 ProfEngel/OwlLM2-e2b 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 ProfEngel/OwlLM2-e2b to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for ProfEngel/OwlLM2-e2b to start chatting
- Docker Model Runner
How to use ProfEngel/OwlLM2-e2b with Docker Model Runner:
docker model run hf.co/ProfEngel/OwlLM2-e2b:Q8_0
- Lemonade
How to use ProfEngel/OwlLM2-e2b with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull ProfEngel/OwlLM2-e2b:Q8_0
Run and chat with the model
lemonade run user.OwlLM2-e2b-Q8_0
List all available models
lemonade list
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# ModelCard – OwlLM2 **Controlling-Experte**
|
| 2 |
|
| 3 |
## Modellübersicht
|
|
@@ -34,7 +42,7 @@ OwlLM2 ist eine speziell auf deutsches Controlling und Rechnungswesen finegetune
|
|
| 34 |
|
| 35 |
| Parameter | Wert |
|
| 36 |
| :-- | :-- |
|
| 37 |
-
| **GPU** | NVIDIA
|
| 38 |
| **Trainingsdauer** | 76,76 Minuten (4.605,80 Sekunden) |
|
| 39 |
| **Peak Memory** | 8,268 GB (37,31% der verfügbaren VRAM) |
|
| 40 |
| **Training Memory** | 0,62 GB (2,80% für LoRA-Training) |
|
|
@@ -84,15 +92,15 @@ OwlLM2 wird in verschiedenen optimierten Formaten bereitgestellt:
|
|
| 84 |
### **Vollständiges Modell (Safetensors)**
|
| 85 |
|
| 86 |
- **Format:** vLLM-kompatible Safetensors
|
| 87 |
-
- **Größe:** ~
|
| 88 |
- **Verwendung:** Direkter Einsatz ohne Basis-Modell
|
| 89 |
- **Ideal für:** Production-Deployments
|
| 90 |
|
| 91 |
|
| 92 |
### **GGUF-Quantisierung**
|
| 93 |
|
| 94 |
-
- **8-Bit GGUF:** ~
|
| 95 |
-
- **4-Bit GGUF:** ~1-2 GB, maximale Kompression
|
| 96 |
- **Ideal für:** Edge-Computing, lokale Anwendungen
|
| 97 |
|
| 98 |
|
|
@@ -319,3 +327,9 @@ Für Fragen, Verbesserungsvorschläge oder spezielle Anwendungsfälle im deutsch
|
|
| 319 |
- Kann Fehler enthalten oder veraltete Informationen wiedergeben
|
| 320 |
|
| 321 |
Entscheidungen, die auf Basis der Modell-Ausgaben getroffen werden, erfolgen auf eigenes Risiko und eigene Verantwortung.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- de
|
| 5 |
+
base_model:
|
| 6 |
+
- google/gemma-3n-E2B-it
|
| 7 |
+
pipeline_tag: text-to-speech
|
| 8 |
+
---
|
| 9 |
# ModelCard – OwlLM2 **Controlling-Experte**
|
| 10 |
|
| 11 |
## Modellübersicht
|
|
|
|
| 42 |
|
| 43 |
| Parameter | Wert |
|
| 44 |
| :-- | :-- |
|
| 45 |
+
| **GPU** | NVIDIA L4 (24GB VRAM) |
|
| 46 |
| **Trainingsdauer** | 76,76 Minuten (4.605,80 Sekunden) |
|
| 47 |
| **Peak Memory** | 8,268 GB (37,31% der verfügbaren VRAM) |
|
| 48 |
| **Training Memory** | 0,62 GB (2,80% für LoRA-Training) |
|
|
|
|
| 92 |
### **Vollständiges Modell (Safetensors)**
|
| 93 |
|
| 94 |
- **Format:** vLLM-kompatible Safetensors
|
| 95 |
+
- **Größe:** ~7-9 GB
|
| 96 |
- **Verwendung:** Direkter Einsatz ohne Basis-Modell
|
| 97 |
- **Ideal für:** Production-Deployments
|
| 98 |
|
| 99 |
|
| 100 |
### **GGUF-Quantisierung**
|
| 101 |
|
| 102 |
+
- **8-Bit GGUF:** ~4-5 GB, optimiert für CPU/kleine GPUs
|
| 103 |
+
- **4-Bit GGUF:** ~1-2 GB, maximale Kompression (noch nicht verfügbar!)
|
| 104 |
- **Ideal für:** Edge-Computing, lokale Anwendungen
|
| 105 |
|
| 106 |
|
|
|
|
| 327 |
- Kann Fehler enthalten oder veraltete Informationen wiedergeben
|
| 328 |
|
| 329 |
Entscheidungen, die auf Basis der Modell-Ausgaben getroffen werden, erfolgen auf eigenes Risiko und eigene Verantwortung.
|
| 330 |
+
|
| 331 |
+
**OwlLM2 – Ihr KI-gestützter Partner für professionelles Controlling**
|
| 332 |
+
|
| 333 |
+
<div style="text-align: center">⁂</div>
|
| 334 |
+
|
| 335 |
+
[^1]: Gemma3N_-4B-_Conversational.ipynb
|