AbdelStark/CodingJEPA

View on GitHub

A Joint-Embedding Predictive Architecture for Python code, based on LeWorldModel architecture and training paradigm.

Stars: 0
Forks: 0
Open beginner issues: 0
Indexed issues: 32

Dominant language: Python
License: No license data
Last GitHub push: May 15, 2026
Latest indexed: Sep 21, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed
PR merge metrics: No merged PRs in 30d

Beginner friendly Everything

Open All

32 open issues indexed

release: create GitHub tag data-v1.0.0 at the manifest commit (RFC-0015) Open

area:release effort:s priority:p1 spec:rfc-0009 spec:rfc-0015 type:docs

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

AbdelStark/CodingJEPA#197 · Opened May 15, 2026
release: verify public HF access — load_dataset() roundtrip smoke test (RFC-0015) Open

area:release effort:s priority:p0 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 68/100

AbdelStark/CodingJEPA#196 · Opened May 15, 2026
release: publish dataset card (README.md) to HF Hub (RFC-0015 Step 13) Open

area:release effort:s priority:p1 spec:rfc-0015 type:docs

Difficulty 2/5 1-3 hours Newbie friendliness 75/100

AbdelStark/CodingJEPA#195 · Opened May 15, 2026
release: upload NOTICE and LICENSES/ to HF dataset repo (RFC-0014 D2–D3) Open

area:release effort:s priority:p1 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 2/5 1-3 hours Newbie friendliness 68/100

AbdelStark/CodingJEPA#194 · Opened May 15, 2026
release: upload chunks config to CodingJEPA/human-python-refactors on HF Hub (RFC-0015) Open

area:release effort:m priority:p1 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 68/100

AbdelStark/CodingJEPA#193 · Opened May 15, 2026
release: upload pairs config to CodingJEPA/human-python-refactors on HF Hub (RFC-0015) Open

area:release effort:m priority:p1 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 65/100

AbdelStark/CodingJEPA#192 · Opened May 15, 2026
release: dry-run hf_convert.py — validate HF Datasets schema on full corpus sample Open

area:release effort:s priority:p1 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 56/100

AbdelStark/CodingJEPA#191 · Opened May 15, 2026
release: write data/DATASET_CARD.md from RFC-0015 template (human-python-refactors) Open

area:release effort:s priority:p1 spec:rfc-0015 type:docs

Difficulty 3/5 1-2 days Newbie friendliness 68/100

AbdelStark/CodingJEPA#190 · Opened May 15, 2026
data: gold subset annotation — 200 pairs (25/intent), tooling + Cohen's κ ≥ 0.7 gate (RFC-0002 D7) Open

area:data effort:l priority:p1 spec:rfc-0002 spec:rfc-0015 type:feature

Difficulty 5/5 Over a week Newbie friendliness 35/100

AbdelStark/CodingJEPA#189 · Opened May 15, 2026
data: corpus QA report — per-repo statistics, intent distribution, pair count validation Open

area:data effort:m priority:p1 spec:rfc-0015 type:docs

Difficulty 3/5 1-2 days Newbie friendliness 72/100

AbdelStark/CodingJEPA#188 · Opened May 15, 2026
data: Step 10 — write content-addressed manifest.lock.json and commit (RFC-0014, RFC-0015) Open

area:data effort:s priority:p0 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 58/100

AbdelStark/CodingJEPA#187 · Opened May 15, 2026
data: Step 9 — full corpus audit, all 4 gates must pass (RFC-0002 D2, RFC-0014 D10) Open

area:data effort:m priority:p0 spec:rfc-0002 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 42/100

AbdelStark/CodingJEPA#186 · Opened May 15, 2026
data: Step 8 — assign train/val/test splits by-repository (RFC-0014 D7) Open

area:data effort:s priority:p0 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 55/100

AbdelStark/CodingJEPA#185 · Opened May 15, 2026
data: Step 7 — deduplication (exact SHA-256 + MinHash LSH near-dedup) (RFC-0014 D6) Open

area:data effort:m priority:p0 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 52/100

AbdelStark/CodingJEPA#184 · Opened May 15, 2026
data: Step 6 — secret scan over pairs corpus, zero-hit gate (RFC-0014 D5) Open

area:data effort:s priority:p0 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 45/100

AbdelStark/CodingJEPA#183 · Opened May 15, 2026
data: Step 5 — apply 8 heuristic labelers + enforce 12k/intent quota (RFC-0002 D6–D8) Open

area:data effort:m priority:p0 spec:rfc-0002 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 45/100

AbdelStark/CodingJEPA#182 · Opened May 15, 2026
data: Step 4 — extract refactor pairs with --cutoff 2023-12-31 (RFC-0002 D3–D4, D11) Open

area:data effort:l priority:p0 spec:rfc-0002 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 42/100

AbdelStark/CodingJEPA#181 · Opened May 15, 2026
data: Step 3 — chunk all repos → data/chunks/*.parquet (RFC-0012) Open

area:data effort:m priority:p0 spec:rfc-0012 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 48/100

AbdelStark/CodingJEPA#180 · Opened May 15, 2026
data: Step 2 — normalize all repo source files (RFC-0012, RFC-0014 D5) Open

area:data effort:m priority:p0 spec:rfc-0012 spec:rfc-0014 spec:rfc-0015 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 55/100

AbdelStark/CodingJEPA#179 · Opened May 15, 2026
data: Step 1 — mirror all 10 OSS repos at pinned SHAs (RFC-0002 D1, RFC-0015 Step 1) Open

area:data effort:m priority:p0 spec:rfc-0002 spec:rfc-0015 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 55/100

AbdelStark/CodingJEPA#178 · Opened May 15, 2026
[Tracking] Dataset: human-python-refactors — curation, quality gates, and HF publication Open

area:data area:release priority:p0 spec:rfc-0002 spec:rfc-0014 spec:rfc-0015 type:tracking

Difficulty 5/5 Over a week Newbie friendliness 25/100

AbdelStark/CodingJEPA#173 · Opened May 15, 2026
release: HF Hub upload script for labeled pairs corpus Open

area:release effort:m priority:p1 spec:rfc-0014 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 48/100

AbdelStark/CodingJEPA#127 · Opened May 15, 2026
release: HF Hub upload script for model weights + tokenizer Open

area:release effort:m priority:p1 spec:rfc-0013 type:feature

Difficulty 3/5 1-2 days Newbie friendliness 68/100

AbdelStark/CodingJEPA#126 · Opened May 15, 2026
docs: PHASE-2.md and PHASE-3.md notes (training observations + departures) Open

area:training effort:s priority:p1 type:docs

Difficulty 3/5 1-2 days Newbie friendliness 55/100

AbdelStark/CodingJEPA#77 · Opened May 15, 2026
train: Stage B intent fine-tune run (50k steps, intent-balanced) Open

area:training effort:l priority:p0 spec:rfc-0008 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 35/100

AbdelStark/CodingJEPA#76 · Opened May 15, 2026
train: Stage A pretrain run (200k steps, B=64, grad_accum=16) Open

area:training effort:l priority:p0 spec:rfc-0008 type:feature

Difficulty 5/5 Over a week Newbie friendliness 25/100

AbdelStark/CodingJEPA#75 · Opened May 15, 2026
infra: Dockerfile.eval pinned by digest Open

area:infra effort:m priority:p1 spec:rfc-0013 type:feature

Difficulty 4/5 3-5 days Newbie friendliness 66/100

AbdelStark/CodingJEPA#24 · Opened May 15, 2026
[Tracking] Paper, release, packaging (paper draft + MODEL_CARD + HF Hub + v1.0 tag) Open

area:release effort:l priority:p0 spec:rfc-0011 spec:rfc-0013 type:tracking

Difficulty 5/5 Over a week Newbie friendliness 25/100

AbdelStark/CodingJEPA#12 · Opened May 15, 2026
[Tracking] Training pipeline (Stage A pretrain + Stage B intent fine-tune) Open

area:training effort:l priority:p0 spec:rfc-0008 type:tracking

Difficulty 5/5 Over a week Newbie friendliness 25/100

AbdelStark/CodingJEPA#6 · Opened May 15, 2026
[Tracking] Data pipeline (mirror, chunk, label, dedup, splits) Open

area:data effort:l priority:p0 spec:rfc-0002 spec:rfc-0012 spec:rfc-0014 type:tracking

Difficulty 5/5 Over a week Newbie friendliness 20/100

AbdelStark/CodingJEPA#4 · Opened May 15, 2026
[Tracking] Infrastructure & reproducibility Open

area:infra effort:l priority:p0 spec:rfc-0013 type:tracking

Difficulty 5/5 Over a week Newbie friendliness 25/100

AbdelStark/CodingJEPA#3 · Opened May 15, 2026
[meta] tracker: spec corpus + implementation issue set bootstrap Open

area:meta effort:s priority:p1 spec:bootstrap type:tracking

Difficulty 4/5 3-5 days Newbie friendliness 35/100

AbdelStark/CodingJEPA#2 · Opened May 15, 2026

AbdelStark/CodingJEPA

Get new issues in your inbox