marauder-actual 4cef9386b1 add docs: system lora plan, specialist specs, training review

2026-05-31 11:38:46 +02:00

bt7274-lora-v1

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

data

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

docker @ d6a6541fb5

add madcat-ml as submodule at ./docker

2026-05-31 11:31:02 +02:00

docs

add docs: system lora plan, specialist specs, training review

2026-05-31 11:38:46 +02:00

review

add docs: system lora plan, specialist specs, training review

2026-05-31 11:38:46 +02:00

.gitattributes

enable git-lfs, track *.png

2026-05-31 11:38:25 +02:00

.gitignore

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

.gitmodules

add madcat-ml as submodule at ./docker

2026-05-31 11:31:02 +02:00

bt7274_memory_100.jsonl

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

bt7274_memory_1000.jsonl

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

bt7274_persona.jsonl

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

bt7274_v1.jsonl

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

bt7274_v2.jsonl

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

bt7274_v3_reformatted.jsonl

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

bt7274_v3.jsonl

v3 dataset (582 ex, 14% direct) + Qwen3.5-27B training script

2026-05-25 18:57:17 +02:00

bt7274_v4.jsonl

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

build_v4.py

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

core_agent_tools.jsonl

add training datasets: memory, persona, agent tools, serpent data

2026-05-31 11:38:37 +02:00

extract_specialists.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

gen_agent_dataset.py

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

gen_memory_dataset_v2.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

gen_memory_dataset.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

gen_persona_dataset.py

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

gen_tts_dataset.py

feat: tts-norm LoRA — dataset generator + training script

2026-05-26 00:14:51 +02:00

gen_v3.py

v3 dataset (582 ex, 14% direct) + Qwen3.5-27B training script

2026-05-25 18:57:17 +02:00

justfile

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

merge.py

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

mine_repos.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

reformat_v3.py

feat: bt7274 LoRA v4 — Hermes format, think blocks, 802 examples

2026-05-26 04:03:38 +02:00

repos.json

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

smoke_test.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

train_memory_lora_v2.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

train_memory_lora.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

train_qwen35_27b.py

fix: parse tool_call arguments from JSON string to dict for Qwen3.5 template

2026-05-25 19:16:07 +02:00

train_specialist.py

add training scripts: memory, specialist, mining, smoke test

2026-05-31 11:38:42 +02:00

train_tts_norm.py

feat: tts-norm LoRA — dataset generator + training script

2026-05-26 00:14:51 +02:00

train_v1.py

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

train_v4.py

fix: Qwen3.5 VL processor has no .encode() — use .tokenizer attr

2026-05-26 14:28:08 +02:00

train.py

init: bt7274 lora training — v1 adapter, v2 dataset (500 examples), justfile

2026-05-25 16:37:30 +02:00

training_data_review.md

add docs: system lora plan, specialist specs, training review

2026-05-31 11:38:46 +02:00

S

Description

LoRA fine-tuning pipeline for building persona, specialist, and voice adapters.

9.9 MiB

Languages

Python 96.7%

Jinja 2%

Just 1.3%