PorTAL for Qwen3-1.7B
This is a native PorTAL artifact for
Qwen/Qwen3-1.7B. It contains a shared 14-task latent table and canonical LoRA-generating core,
jointly trained with Qwen3-4B, plus the alignment specific to Qwen3-1.7B. It generates rank-8 LoRA
factors for the query and value projections of every decoder layer.
Evaluation
One seed was evaluated on the complete 14-task validation suite using continuation log-probability
divided by character length (acc_norm). Gold continuation token-mean NLL was tracked separately
for checkpoint selection.
| Model | Macro acc_norm |
|---|---|
| Frozen Qwen3-1.7B | 0.5745 |
| PorTAL-adapted | 0.7096 |
| Absolute lift | +0.1351 |
These are research benchmark results for this exact artifact and evaluation recipe, not a general performance guarantee.
Training recipe
- Base:
Qwen/Qwen3-1.7Bat70d244cc86ccca08cf5af4e1e306ecf908b1ad5e - Joint source bases: Qwen3-1.7B and Qwen3-4B
- Dataset:
RampPublic/portallib-tasksatffc3c0e44f529bf64a5ae62ed5db090952db97ea - Source data: leading 2,000 examples per task, or all available examples for shorter tasks; fixed across epochs
- Optimization: 500 balanced rounds per epoch, 12 epochs, batch size 4, core/alignment LR
1e-3, latent LR2e-3, linear decay with 10% warmup, seed 0 - Checkpoint: maximum macro validation
acc_norm, with lower gold NLL as the tie-breaker - Architecture: q/v targets, rank 8, alpha 16, task latent 256, layer embedding 32, hidden 512, canonical width 1024
The Qwen3-1.7B and Qwen3-4B source artifacts contain identical shared task latents and canonical core weights; only their base-specific alignments differ.
Usage
from portallib import PortalModel
portal = PortalModel.from_pretrained(
"RampPublic/portal-qwen3-1.7b",
revision="v0.2.0",
)
portal.export_peft("rte", "./portal-rte-qwen3-1.7b")
See the release recipe for the full task list, evaluation definition, and training procedure. The artifact is Apache-2.0; the benchmark dataset contains components under multiple upstream licenses documented on its dataset card.