Instructions to use od2961/maxent-grpo-models with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use od2961/maxent-grpo-models with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("od2961/maxent-grpo-models", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download experiments/E98-R1/README.md from od2961/maxent-grpo-models: direct link, hf CLI and curl.
- Browser
- Download file 11.1 kB
-
https://huggingface.co/od2961/maxent-grpo-models/resolve/main/experiments/E98-R1/README.md
- Command line
-
hf download hf://od2961/maxent-grpo-models/experiments/E98-R1/README.md
-
curl -L -o README.md https://huggingface.co/od2961/maxent-grpo-models/resolve/main/experiments/E98-R1/README.md
E98-R1 · Sparse RLEP-Dr · Qwen2.5-0.5B
Which aspect of the update changes correctness and retained verified support?
Archive home · Study design · Coverage · Models · Comparators
Study design
Corrected sparse-success-replay recipe for Graph Coloring, Python Factors and PantryPlan, five seeds each.
Direct comparator family. Apply the paper’s frozen model/domain/seed inclusion rules; scientific record counts and verified downloadable model counts are separate.
These are terminal exports from the registered eight-pass training horizon. The stored export-step label is 3073; the papers’ audited pass-8 evaluation endpoint is optimizer step 3072. The archive preserves that original export convention and does not imply an additional evaluated training update.
Archive coverage
| Registered new training cells | Scientific records selected | Deployable exports selected | Verified exports |
|---|---|---|---|
| 15 | 15 | 15 | 15 |
Scientific records preserve the admitted result identity. Deployable exports additionally have available model weights; verified exports have completed remote verification. Cells outside the fixed selection are not silently counted, and unverified model folders are not linked below.
| Model | Verified / selected | Seeds in selection |
|---|---|---|
| Qwen2.5-0.5B-Instruct | 15 / 15 | 43, 44, 45, 46, 47 |
Methods and collections
Sparse RLEP-Dr · rlep_dr_sparse
Sparse replay of verified successes in observed frequency, without canonical-key balancing.
- Qwen2.5-0.5B-Instruct · Graph Coloring — 5 verified exports. Exact immutable revisions are in the model tables below.
- Qwen2.5-0.5B-Instruct · PantryPlan — 5 verified exports. Exact immutable revisions are in the model tables below.
- Qwen2.5-0.5B-Instruct · Python Factors — 5 verified exports. Exact immutable revisions are in the model tables below.
Verified models
Browse by domain: Countdown · Graph Coloring · Python Factors · MathIR · PantryPlan.
Countdown
0 / 0 selected exports verified.
Transfer in progress. No verified model folders are available in this subset yet.
Graph Coloring
5 / 5 selected exports verified.
| Model | Method | Seed | Terminal export | Immutable revision |
|---|---|---|---|---|
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 43 | step 3073 | 5b6a02106fb0 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 44 | step 3073 | a824988e9188 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 45 | step 3073 | 4722df20b7c6 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 46 | step 3073 | 8db4f77a2105 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 47 | step 3073 | cbd44a9f5459 |
Python Factors
5 / 5 selected exports verified.
| Model | Method | Seed | Terminal export | Immutable revision |
|---|---|---|---|---|
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 43 | step 3073 | 1a4f9a67d364 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 44 | step 3073 | 5aad6555ef42 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 45 | step 3073 | 1d29ce5219f8 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 46 | step 3073 | 36ad376062f8 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 47 | step 3073 | 84d3f195443a |
MathIR
0 / 0 selected exports verified.
Transfer in progress. No verified model folders are available in this subset yet.
PantryPlan
5 / 5 selected exports verified.
| Model | Method | Seed | Terminal export | Immutable revision |
|---|---|---|---|---|
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 43 | step 3073 | 04b362929856 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 44 | step 3073 | 763cf3c17696 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 45 | step 3073 | d5551ca79dac |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 46 | step 3073 | 80e95b09a6f1 |
| Qwen2.5-0.5B-Instruct | Sparse RLEP-Dr | 47 | step 3073 | 27df834cc578 |
Model-family shortcuts
Qwen2.5-0.5B-Instruct
Browse verified collections. Domain tables above provide exact model revisions.
Comparators and analysis scope
Use E78 references; E116 supplies further domains. Preserve the corrected E98-R1 source identity rather than relabeling it as superseded E98.
Restore and verify a model · Applicable base-model licenses. The full commit_sha, source identity and file provenance are available through the root catalog and each pinned export manifest.