Не попадитесь на накрученные каналы! Узнайте, не накручивает ли канал просмотры или
подписчиков
Проверить канал на накрутку
Телеграм канал «Нейро Примус»
Нейро Примус
2.1K
0
3
0
2.4K
Нейронные сети — в каждый дом, креатив в каждую голову!
Подписчики
Всего
454
Сегодня
0
Просмотров на пост
Всего
64
ER
Общий
13.99%
Суточный
10.57%
Динамика публикаций
Telemetr - сервис глубокой аналитики
телеграм-каналов
телеграм-каналов
Получите подробную информацию о каждом канале
Отберите самые эффективные каналы для
рекламных размещений, по приросту подписчиков,
ER, количеству просмотров на пост и другим метрикам
рекламных размещений, по приросту подписчиков,
ER, количеству просмотров на пост и другим метрикам
Анализируйте рекламные посты
и креативы
и креативы
Узнайте какие посты лучше сработали,
а какие хуже, даже если их давно удалили
а какие хуже, даже если их давно удалили
Оценивайте эффективность тематики и контента
Узнайте, какую тематику лучше не рекламировать
на канале, а какая зайдет на ура
на канале, а какая зайдет на ура
Показано 7 из 2 111 постов
Смотреть все посты
Пост от 24.10.2026 07:39
35
0
3
ComfyUI Character Card Studio
This node takes one reference image or a text prompt, generates every angle, expression, and detail sheet through Qwen-Image 2.1, and slices the output straight into preset character sheets without cropping feet or heads.
👉 GitHub
This node takes one reference image or a text prompt, generates every angle, expression, and detail sheet through Qwen-Image 2.1, and slices the output straight into preset character sheets without cropping feet or heads.
👉 GitHub
Пост от 24.10.2026 06:38
2
0
0
ComfyUI Speech-Bubble
A post-process node that overlays readable manga speech bubbles and text on generated images
👉 GitHub
A post-process node that overlays readable manga speech bubbles and text on generated images
👉 GitHub
Пост от 24.10.2026 01:25
5
0
1
MiniMax-H3 PTD + RefAsymOPD Ref2VA Acceleration LoRA
It works well for what it targets.
LoRA fixes the main issue with straight PTD, which tends to smudge background details when you push for speed.
By keeping the asymmetric distillation on the reference path, it forces the model to hold onto texture without adding inference latency.
And 8 steps is a sweet spot for local runs
👉 HF
It works well for what it targets.
LoRA fixes the main issue with straight PTD, which tends to smudge background details when you push for speed.
By keeping the asymmetric distillation on the reference path, it forces the model to hold onto texture without adding inference latency.
And 8 steps is a sweet spot for local runs
👉 HF
Пост от 24.10.2026 00:02
13
0
0
Qwen-Image 2.1 VAE 8px fix
Fine-tune of Texture-Fix that fades the 8-pixel grid on fur.
Works well, especially for high‑res images.
👉 HF
👉 HF Comfyui
Fine-tune of Texture-Fix that fades the 8-pixel grid on fur.
Works well, especially for high‑res images.
👉 HF
👉 HF Comfyui
Пост от 23.10.2026 23:23
10
0
0
QI-2.1 Pokemon Foil Mask LoRA
Hmm… not sure what this is actually for .. but: drop in any Pokemon-style card, get back its foil mask
👉 HF
Hmm… not sure what this is actually for .. but: drop in any Pokemon-style card, get back its foil mask
👉 HF
Пост от 20.10.2026 20:32
29
0
1
ComfyUI-LUT Master
LUT post-processing node for ComfyUI.
👉 GitHub
LUT post-processing node for ComfyUI.
👉 GitHub
👍
1
Пост от 20.10.2026 14:14
42
0
0
FLUX 3 Max
Artificial Analysis makes some tests...
Flux3 is fast and cheap for what it does. Ten seconds rendered in under twelve seconds at ten dollars twenty a cent a minute, dropping to two fifty five during the launch discount (on fal).
Wins on physics and live action ads. But...fumbles camera control and human anatomy (also their iconic chin).
If you need a quick product shot with good lighting, use it. If you need characters with hands that make sense, wait for the next patch.. or other model.
Anyway, H3 is still the best local model for pretty much anything.
#ainews
Artificial Analysis makes some tests...
Flux3 is fast and cheap for what it does. Ten seconds rendered in under twelve seconds at ten dollars twenty a cent a minute, dropping to two fifty five during the launch discount (on fal).
Wins on physics and live action ads. But...fumbles camera control and human anatomy (also their iconic chin).
If you need a quick product shot with good lighting, use it. If you need characters with hands that make sense, wait for the next patch.. or other model.
Anyway, H3 is still the best local model for pretty much anything.
#ainews
👍
1