Downloads · 30 days
11
7% of all-time downloads
parser-qwen3-0.6b
powerliftme/parser-qwen3-0.6b
parser-qwen3-0.6b is a text generation model from powerliftme. Use it when you need the model to write or continue text. It is set up for gguf. The card lists the license as apache-2.0.
On-device parser for the PowerliftME app. Turns a free-text workout log into a canonical, language-neutral JSON record (movement family, sets, reps, load, load %, RPE/RIR, role), grammar-constrained at decode time.
All-time downloads
165
Public
Repo size
639 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.gguf639 MB · 100%
At a glance
- Task
- Text Generation
- Library
- gguf
- License
- apache-2.0
- Access
- Public
- Created
- Jun 14, 2026
- Updated
- Jun 14, 2026
- SHA
- 43ce85d0
Try a prompt
Base models
Model card
From the Hugging Face model README
PowerliftME · Workout Parser (Qwen3 0.6B, GGUF)
On-device parser for the PowerliftME app. Turns a free-text workout log into a canonical, language-neutral JSON record (movement family, sets, reps, load, load %, RPE/RIR, role), grammar-constrained at decode time.
- Base: Qwen3-0.6B (Apache 2.0)
- Quant: Q8_0 · ~0.64 GB
- Languages: English + Russian input; canonical output
- Held-out field accuracy: EN 96.3% · RU 92.7%
Run (llama.cpp)
llama-server -m parser-qwen3-0.6b-q8.gguf -c 1024
Feed the app's parser system prompt plus the JSON grammar; the model emits one
canonical record per exercise. Numbers come from the input only — empty load
stays absent (sentinel), an explicit 0 is a real zero.
License
Apache 2.0, inherited from Qwen3-0.6B. No usage restrictions.
- Task
- Text Generation
- Library
- gguf
- License
- apache-2.0
- Languages
- en, ru
- Created
- Jun 14, 2026
- Updated
- Jun 14, 2026