Changlin LI

Computer Vision Engineering

[Changlin LI @ChanglinLI1]

gpt 5.6 + link ibkr后真好用

[Changlin LI @ChanglinLI1]

@sth_like_human @Aether8964 @HuoQubot

[Changlin LI @ChanglinLI1]

@MiaBleem 新粉报道

[AK @_akhaliq]

Tencent just released Hunyuan Video

A Systematic Framework For Large Video Generation Model Training

[汗青 HQ @hq4ai]

提前拿到腾讯混元视频模型的内测资格，发现还是有挺多亮点，文生视频的几个样片：
1️⃣物理特性还原度很高，没出现太多错误开卡，马匹奔腾，海底生物、松鼠的动态生动合理
2️⃣文生视频的美学表也不错，尤其西部片电影这支样片感觉很到位。
混元也开始发力了，期待一波后续更新和测试。
样片1👇

[Vaibhav (VB) Srivastav @reach_vb]

Mixtral 8x7B Instruct with AWQ & Flash Attention 2 🔥

All in ~24GB GPU VRAM!

With the latest release of AutoAWQ - you can now run Mixtral 8x7B MoE with Flash Attention 2 for blazingly fast inference.

All in < 10 lines of code.

The only real change except loading AWQ weights is to pass attn_implementation="flash_attention_2" over to the .from_pretrained call whilst loading the model.

Here's a full run through:

1. Install AutoAWQ and transformers

pip install autoawq git+https://github. com/huggingface/transformers.git

2. Initialise the tokeniser and the model

from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer

model_id = "casperhansen/mixtral-instruct-awq"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
low_cpu_mem_usage=True,
device_map="cuda:0",
attn_implementation="flash_attention_2")

3. Initialise the TextStreamer

streamer = TextStreamer(tokenizer,
skip_prompt=True,
skip_special_tokens=True)

4. Tokenise the inputs

tokens = tokenizer(
text,
return_tensors='pt'
).input_ids.to("cuda:0")

5. Generate!

generation_output = model.generate(
tokens,
streamer=streamer,
max_new_tokens=512
)

That's it! 🤗

[Changlin LI @ChanglinLI1]

How many ram will the finetuning cost? Is 16G ram mbp available?

[Awni Hannun @awnihannun]

Fine-tuning Mistral 7B with LoRA on a 32 GB M1 (laptop!) in MLX

https://t.co/BCQZAWHCTA

Updated example uses less RAM + support for custom datasets 🚀

[Changlin LI @ChanglinLI1]

Recreate apple's logo with stylar's new logo feature @stylar_ai
#AI #AIArtwork #AIphoto

[MindBranches @MindBranches]

Link to OpenAI’s full Prompt Engineering Guide:
https://t.co/xhNBs9niQx

[Changlin LI @ChanglinLI1]

@GeorgeGCanada1 @Jacob43817149 确实有点像啊@wangwatchworld

[Changlin LI @ChanglinLI1]

Use @stylar_ai to generate Flying Spaghetti Monster, so cute!🤩
#AIart #ai #designwithstylar

[Dzine @dzine_ai]

Welcome to Stylar! 😊 Here's our beginner's guide video to get you started on your creative journey. 🎨

[Changlin LI @ChanglinLI1]

Use @stylar_ai to give Nami more variation 🤗
#AIArtworks #aiart #waifu #Nami #designwithstylar #digitalart

[Nami ナミ 👑 @NamiTrueQueen]

Beautiful Nami ♥️
Credit to : bonnieaiart (DeviantArt).

#ONEPIECE #ONEPIECE1099 #ワンピース #ルフィ #ナミ #luffy #nami #waifu #aiart #Twitter

[Changlin LI @ChanglinLI1]

@stylar_ai Amazing!

[Changlin LI @ChanglinLI1]

Try to restyle the amazing image using @stylar_ai 🤗
#stylar #AIart #digitalart

[astro @ainextastro]

𝖥𝗋𝗂𝖽𝖺𝗒
#midjourney

[sundyme @sundyme]

分享一个非常漂亮的 Midjourney 风格，特别适合东方人物形象和服饰。这个风格对出图影响很明显，画面明亮通透，色彩艳丽，细节细腻。
风格代码是：cmEX5159uqALTIBd 值得一提的是生成这个风格只选择了3个选项（1左，3左，32左），还是很神奇的。关于 Midjourney 风格定制的一些有趣���现，1/2

[Changlin LI @ChanglinLI1]

I use Stylar ai @stylar_ai to restyle those images, really amazing!

#Halloween #Halloween2023 #halloweenmakeup #digitalart #digitaldrawing #ai #designwithstylar #stylarai

[ArtiMindArt @artimindArt]

🚨 HALLOWEEN PROMPT ALERT!

Dive into the spooky carnival of your nightmares with Spookytoon Halloween Art Magic! 🎃👻 Discover eerie witches, menacing ghosts, and more in this mysterious world of art. Don't miss out on the spine-tingling adventure!

/imagine Scary carnival of [monster/object...] as Spookytoon Halloween Art Magic --s 200

Halloween is coming...I would love to see your creations.

I leave you in the alts some examples!

#AIart #AIArtCommuity #promptshare

[Dzine @dzine_ai]

We found it hard to prompt out desired styles in Gen-AI tools. That's why we came up with a curated styles library in Stylar for our fellow designers, specially selected to cater to different commercial uses. 🎯

Try Stylar Beta: https://t.co/SqHyKE3zK4

#MadeInStylar

[Dzine @dzine_ai]

Coming soon!
Introducing Stylar (Beta): an AI-powered design partner for creative minds. ✍🎨
Follow for updates^ #StylarAI

[Changlin LI @ChanglinLI1]

@wangwatchworld 上学时候认识的韩国人普遍英文水平高国人很多