Hugging Face's logo
zzl's picture
6 7

zzl

zzlxv
·

AI & ML interests

None yet

Recent Activity

reacted to fdaudens's post with 👀 19 days ago
🎯 Perplexity drops their FIRST open-weight model on Hugging Face: A decensored DeepSeek-R1 with full reasoning capabilities. Tested on 1000+ examples for unbiased responses. Check it out: https://huggingface.co./perplexity-ai/r1-1776 Blog post: https://perplexity.ai/hub/blog/open-sourcing-r1-1776
upvoted a paper 22 days ago
3CAD: A Large-Scale Real-World 3C Product Dataset for Unsupervised Anomaly
upvoted a paper 26 days ago
The Curse of Depth in Large Language Models
View all activity

Organizations

None yet

zzlxv's activity

reacted to fdaudens's post with 👀 19 days ago
upvoted a paper 22 days ago
upvoted a paper 26 days ago
upvoted an article about 1 month ago
view article
Article

You could have designed state of the art positional encoding

• 179
upvoted a paper about 1 month ago
liked a Space 3 months ago
reacted to merve's post with 👀 3 months ago
upvoted a paper 3 months ago
reacted to ImranzamanML's post with 👍 5 months ago
view post
Post
2753
Here is how we can calculate the size of any LLM model:

Each parameter in LLM models is typically stored as a floating-point number. The size of each parameter in bytes depends on the precision.

32-bit precision: Each parameter takes 4 bytes.
16-bit precision: Each parameter takes 2 bytes

To calculate the total memory usage of the model:
Memory usage (in bytes) = No. of Parameters × Size of Each Parameter

For example:
32-bit Precision (FP32)
In 32-bit floating-point precision, each parameter takes 4 bytes.
Memory usage in bytes = 1 billion parameters × 4 bytes
1,000,000,000 × 4 = 4,000,000,000 bytes
In gigabytes: ≈ 3.73 GB

16-bit Precision (FP16)
In 16-bit floating-point precision, each parameter takes 2 bytes.
Memory usage in bytes = 1 billion parameters × 2 bytes
1,000,000,000 × 2 = 2,000,000,000 bytes
In gigabytes: ≈ 1.86 GB

It depends on whether you use 32-bit or 16-bit precision, a model with 1 billion parameters would use approximately 3.73 GB or 1.86 GB of memory, respectively.
upvoted a paper 5 months ago
reacted to designermohr's post with 👀 8 months ago
view post
Post
566
Is a full-featured Mac sufficient for AI and ML development compared to NVIDIA GPU systems?

Hello everyone,
We are about to make the decision to purchase a powerful Mac with maximum features for our university. This Mac will primarily serve as a development computer in the field of artificial intelligence (AI) and machine learning (ML) and will be used by a small group of users in the local network. After development, the systems will be transferred to servers that have to withstand higher loads and visitor numbers.

Planned system:
Apple Mac Studio 2023 with M2 Ultra processor
24-core CPU
76-core GPU
32-core NPU (neural engine) for machine learning
128GB RAM
1TB HDD

Our question to you:
Is a fully equipped Mac with the latest SoCs chips and integrated neural engines sufficient for the development of AI and ML systems?
Or should we rather rely on proven Windows/Linux systems with powerful NVIDIA graphics cards?

We already have several NVIDIA graphics cards available at the university:
NVIDIA Tesla T4
NVIDIA 2080Ti
NVIDIA 3080Ti

We are particularly interested in your experiences and assessments of how the performance of the Mac compares to the GPU systems mentioned.

Are there significant differences, especially in the development and training of models?

What difference would you consider to be significant?

For us, a difference of 100% or more would be considered significant.
In other words, computer A (Mac) takes twice as long to calculate as computer B (NVIDIA system).

Many thanks in advance for your answers and experience!
Best regards,

Oliver
liked a Space about 1 year ago
liked 3 models over 1 year ago
liked 2 Spaces over 1 year ago