Declare Apache-2.0 licence

pipeline: pass explicit --with deps to child jobs (--no-project suppressed PEP 723 and killed the merge in 3s)

README: document the escaped-newline defect, the BOM dependency bug, and point at HANDOFF.md

Handoff: fixed supervisor + full findings/state so work can resume

Handoff: fixed supervisor + full findings/state so work can resume

Fix: repair JSON-escaped newlines in training data (v1/v2 scored 0% ast_rate vs base 93%)

Bisect: base+adapter vs merged, to locate the escaped-newline defect

Diagnostic: dump full replies to diagnose v2 0% ast_rate

Strip UTF-8 BOM so uv PEP 723 metadata is detected (deps were never installed)

Fix eval harness: honour --max-new-tokens, exclude truncation from syntax score, fall back across dataset splits

Declare torch (and transformers for GGUF) so uv run works

Declare torch (and transformers for GGUF) so uv run works

Declare torch (and transformers for GGUF) so uv run works

Pipeline supervisor v2: stdlib only (no inline script header)

Fix pipeline supervisor: strip UTF-8 BOM that broke the inline script header

Add pipeline supervisor script (waits for v2 training, chains merge then GGUF)

Push adapter at every checkpoint (--push-checkpoints) so cancel/timeout cannot lose a run

JSONL fallback for datasets without Json feature type

Add malware analysis + counter/remediation sources

Resume adapter + coding/security-heavy mix

Eval v8: restore out.append in code-sanity loop

Eval v7: code-sanity try-block covers chat template

Eval v6: restore _fetch_first_rows

Eval v5: BitsAndBytesConfig for MoE 4-bit

Eval v4: --merged-repo, eos_token_id, deps without unsloth

Eval v4: Hermes JSON-string tools, merged-repo loader, eos_token_id

Eval v3: Hermes conv support + longest-block AST

Eval v2: clean rewrite with PEP 723 deps + inlined helpers

Eval: inline helpers (no cross-script import)

Quantise: apt-install cmake before building llama-quantize

Fix PEP 723 blank line

Fix PEP 723 blank line

Quantise safetensors -> GGUF (llama.cpp path)

Merge only (GGUF in separate job)

Add eval script (tool-call validity, code sanity, security MCQ)

Add merge+GGUF script

Option B: --mix-scale, trackio-before-config fix

SecureCoder runbook

Robust Hub push + throughput reporting

Fix MoE router LoRA target + attention-only fallback

Drop dpevzner (config churn); mix ready

Format-adaptive tool rendering + resilient dataset loading

Fix xLAM tool schemas + dpevzner config

SecureCoder trainer v1

initial commit