This model was trained on a large reasoning dataset derived from Pony Alpha (an early checkpoint of GLM-5).
🧬 Datasets
TeichAI/Pony-Alpha-15k
🏗 Base Model
LiquidAI/LFM2.5-1.2B-Thinking
⚡ Use Cases
- Coding
- Science
- Deep research
∑ Dataset Stats
- Cost: $0 USD
- Total tokens (input + output): 43.3M
Sampling Parameters
Liquid AI recommends the following sampling parameters:
| Setting | Value |
|---|---|
| temperature | 0.05 |
| top_k | 50 |
| repeat_penalty | 1.05 |
This LFM-2.5 model was trained 2× faster using Unsloth and Hugging Face's TRL library.
- Downloads last month
- 133