packs

feat: estilo editorial de la IA configurable (editorial-style.php)

#27

by hermes was merged 2026-09-13 00:12:54 +02:00

main

feat/editorial-style

3 / 3

feat: pack wordpress con IA nativa de WP 7.1 — fuera Open WebUI

#26

by hermes was merged 2026-09-12 23:49:12 +02:00

main

feat/wp-native-ai

4 / 4

fix: limite de CPU 6.0 para ollama

#25

by hermes was merged 2026-09-12 23:07:47 +02:00

main

fix/cpu-limit

fix: think=false en Open WebUI (el SYSTEM /no_think no basta)

#24

by hermes was merged 2026-09-12 23:06:36 +02:00

main

fix/disable-thinking

fix: desactivar el thinking de Qwen3 (SYSTEM /no_think)

#23

by hermes was merged 2026-09-12 22:38:20 +02:00

main

fix/no-think

feat: remove.ps1 / remove.sh para desinstalar cada pack

#22

by hermes was merged 2026-09-12 22:25:08 +02:00

main

feat/remove-script

feat: precalentar el modelo al terminar el deploy

#21

by hermes was merged 2026-09-12 22:15:21 +02:00

main

feat/model-warmup

feat: modelo por defecto qwen3:4b (llama3.2 3B enredaba con las tools)

#20

by hermes was merged 2026-09-12 22:05:34 +02:00

main

feat/model-qwen3-4b

fix: sin loops de tool-calls en el chat

#19

by hermes was merged 2026-09-12 21:54:16 +02:00

main

fix/no-tool-loops

feat: feedback del down/rm inicial

#18

by hermes was merged 2026-09-12 21:48:32 +02:00

main

fix/down-feedback

fix: skip de la descarga si el modelo ya esta en el volumen

#17

by hermes was merged 2026-09-12 21:44:35 +02:00

main

fix/skip-model-download

feat: deteccion de GPU NVIDIA y aceleracion automatica

#16

by hermes was merged 2026-09-12 21:33:44 +02:00

main

feat/gpu-detect

fix: limite de RAM 6G para ollama

#15

by hermes was merged 2026-09-12 21:32:12 +02:00

main

fix/ollama-mem-limit

fix: mensaje del wait-loop de Ollama (no duplicaba el del modelo)

#14

by hermes was merged 2026-09-12 21:27:31 +02:00

main

fix/waitloop-label

fix: down/rm ANTES de detectar puertos

#13

by hermes was merged 2026-09-12 21:22:57 +02:00

main

fix/down-before-ports

fix: progreso del download GGUF y 1 solo mensaje de modelo

#12

by hermes was merged 2026-09-12 21:21:24 +02:00

main

fix/pull-progress

fix: num_ctx 16384 en el modelo (el 4096 da 400 en Open WebUI)

#11

by hermes was merged 2026-09-12 21:19:30 +02:00

main

fix/model-numctx

feat: modelo via HF primero, registry de Ollama como ultimo recurso

#10

by hermes was merged 2026-09-12 21:11:03 +02:00

main

fix/model-hf-first

feat: 2o mirror de HF (unsloth) en el fallback del modelo

#9

by hermes was merged 2026-09-12 21:05:06 +02:00

main

fix/hf-second-mirror

feat: fallback del modelo via HuggingFace si el registry falla

#8

by hermes was merged 2026-09-12 21:04:18 +02:00

main

fix/pull-hf-fallback