Downloads · 30 days
0
davidanugraha/R-Few-Qwen3.5-35B-A3B-SWE-Smith-r18
R-Few-Qwen3.5-35B-A3B-SWE-Smith-r18 is a text generation model from davidanugraha. Use it when you need the model to write or continue text. It is set up for peft.
This public repository is a portability snapshot of the separate R-Few Challenger and Solver checkpoint lineages from rfew-qwen35-35b-all198-counter-fix-20260904-r18.
Downloads · 30 days
0
Access
Public
Updated Sep 11, 2026
Repo size
138 GB
Likes
0
Public
Click a slice to open those files.
.pt103 GB · 75%
From the Hugging Face model README
This public repository is a portability snapshot of the separate R-Few
Challenger and Solver checkpoint lineages from
rfew-qwen35-35b-all198-counter-fix-20260904-r18.
It is not a completed R-Few effectiveness claim. Cycle 7 completed without an optimizer update and its terminal data is included. Skipped cycles do not create fake checkpoint numbers: checkpoint steps count successful optimizer updates, not outer-loop cycles.
| Outer cycle producing update | Challenger | Solver | Result |
|---|---|---|---|
| 0 | step 1 | step 1 | both updated |
| 1 | step 2 | unchanged at step 1 | Challenger only |
| 2 | unchanged at step 2 | unchanged at step 1 | no valid synthetic curriculum |
| 3 | step 3 | unchanged at step 1 | Solver selection was synthetic-only |
| 4 | step 4 | step 2 | mixed human/synthetic Solver update |
| 5 | unchanged at step 4 | unchanged at step 2 | zero Challenger variance; human-only Solver selection |
| 6 | unchanged at step 4 | unchanged at step 2 | zero Challenger variance; human-only Solver selection |
| 7 | unchanged at step 4 | unchanged at step 2 | zero valid synthetic tasks; both updates skipped |
Current continuation parents in this snapshot are Challenger step 4 and Solver step 2. Cycle 7 attempted 32 Challenger groups: 30 rectangular groups yielded 240 trajectories, while two groups were discarded after workspace-contamination errors. No valid synthetic task was produced. The Challenger replay had no within-group reward variance, and no mixed human/synthetic Solver replay was emitted.
checkpoints/challenger/step-1/
checkpoints/challenger/step-2/
checkpoints/challenger/step-3/
checkpoints/challenger/step-4/
checkpoints/solver/step-1/
checkpoints/solver/step-2/
metadata/configs/
metadata/docs/
data/cycle-0007/challenger-groups/
data/cycle-0007/backbone/
data/cycle-0007/diagnostics/
data/cycle-0007/resume_cycle7_summary.json
Each checkpoint directory contains the serving LoRA adapter, committed receipt, diagnostic metrics, and the compact FSDP continuation checkpoint with optimizer, scheduler, and RNG state. The Qwen base-model weights are not included.
Qwen3.5-35B-A3B at immutable revision
59d61f3ce65a6d9863b86d2e96597125219dc754.1e-5, 114,688-token trainable context.1e-5, 65,536-token context.[0.2, 0.7], which admits 2--5 successes at K=8.The canonical source manifest contains 4,549 eligible SWE-Smith tasks, but the r18 mechanics config materializes only five unique human tasks total from 16 validated candidates. Those five are reused for Challenger conditioning and are the entire human side of every Solver difficulty pass.
This should not be confused with the paper's separate 0--5 per-prompt anchor sampling rule. Consequently, r18 is valid mechanics evidence but not the final full-human-budget comparison. A future comparison must define a sealed 4,549-task pool, sample only 0--5 of that pool into each Challenger prompt, and separately specify a fixed-budget human Solver schedule comparable across R-Few, Basic GRPO, and the proposed method.
The current implementation also withholds a Solver update unless both human and synthetic tasks survive in the same cycle. This is a conservative SWE adaptation, not a rule inherited from released R-Zero. Released R-Zero instead filters a variable task pool, trains fixed-size batches, drops incomplete final batches, and has no human/synthetic composition constraint.
This repository preserves model continuation state and compact diagnostics. The
larger campaign directory from Cycles 0--6 is intentionally archived separately
to S3. Cycle 7's sealed Challenger groups, backbone, replay, audits, and summary
are included here; repository workspaces and unrelated raw campaign artifacts
are not. Modal also retains the original training artifact prefixes. See
metadata/docs/MACHINE_MIGRATION.md for locations, hashes, and recovery steps.
The campaign ran from an uncommitted working tree. The current checkout changed after Cycle 7 launched, so this checkpoint release is not a substitute for the curated Git commit described in the migration handoff. Do not claim exact source reproduction until that source closure is committed.