Liquid d1
Collection
Liquid AI d1 decision models (d1-3B, d1-omni-600M) in GGUF: measured AD quants and the metrics behind them. • 4 items • Updated
Dataset Viewer
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
d1-3B GGUF, everything behind the numbers
These are the measurements and inputs behind AtomicChat/d1-3B-GGUF.
| Path | What it is |
|---|---|
results.json |
Every measured file, ours and Liquid's, with its size and two sets of numbers. Decision fidelity against BF16: same answer, flips, option KL and total variation distance. Text: KLD and top-1 over eval/neutral |
imatrix/imatrix.gguf, imatrix/calib.txt |
The importance matrix and its 1.5M-token corpus of decision prompts |
decisions/evalset.jsonl |
The 1,285 held-out decisions: token ids and the option token groups |
decisions/*.json |
The option probabilities of every file, the PyTorch reference (ref_torch.json) and the comparisons |
kld/*.kld.log |
llama-perplexity --kl-divergence logs per file |
logs/ |
Conversion, the tensor comparison with Liquid's GGUFs, imatrix, every quantize run, the llama.cpp commit |
scripts/ |
The prompt and corpus builder, the decision harness, the run script, the PyTorch reference |
- Downloads last month
- 362