GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
flyfesan
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
flyfesan
Model Tree
openai-community/gpt2
Input Modalities
Output Modalities
Supported Functionality
Generate CrossFit-style workouts by prompting with an opening line (e.g. amrap 20 minutes) and sampling (top_k=50, temperature=0.9). The model reproduces the dataset's style and vocabulary, including its quirks (e.g. dumbell), not necessarily real-world programming wisdom. Small dataset (~136K tokens), so output can be repetitive.
Full train split of flyfesan/wods (29,943 rows → 136,142 tokens → 1,063 blocks of 128 tokens); 95/5 train/eval split, seed 42.
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.4852 | 1.5625 | 50 | 1.3917 |
| 1.3315 | 3.0 | 96 | 1.2975 |
Training was run on Hugging Face Jobs (t4-small, ~2.5 min); the live loss curve is on the Trackio dashboard.
Prompt: amrap 20 minutes
text
amrap 20 minutes40 air squat40 dumbell kettlebell swing40 air squat30 dumbell kettlebell swing30 air squat30 dumbell kettlebell swing30 air squat30 dumbell kettlebell swing20 air squat10 dumbell kettlebell swing10 air squatamrap 20 minutes40 air squat40 dumbell kettlebell swing40 air squat30 dumbell kettlebell swing30 air squat30 dumbell kettlebell swing30 air squat30 dumbell kettlebell swing20 air squat10 dumbell kettlebell swing10 air squat