Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
sahil_chaudhary
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
4 ms
·
1.
▲
Predicting and Altering Human Opinions
(hunchfox.com)
3 points
by
sahil_chaudhary
24d ago
|
0 comments
2.
▲
Forecasting the AI Boom till 2030
(hunchfox.substack.com)
3 points
by
sahil_chaudhary
25d ago
|
0 comments
3.
▲
We need to build more data centers
(hunchfox.substack.com)
3 points
by
sahil_chaudhary
26d ago
|
1 comments
4.
▲
What happens if we build less AI infrastructure vs. more: a 2027-2030 forecast
(hunchfox.substack.com)
3 points
by
sahil_chaudhary
27d ago
|
0 comments
5.
▲
Factorio Automation Mod for AI Agents
(github.com)
6 points
by
sahil_chaudhary
2y ago
|
0 comments
6.
▲
Benchmarking code LLMs using user preference
(arena.glaive.ai)
2 points
by
sahil_chaudhary
3y ago
|
1 comments
7.
▲
Code Models Arena: side-by-side battles between code language models
(arena.glaive.ai)
1 points
by
sahil_chaudhary
3y ago
|
0 comments
8.
▲
Open sourcing OpenAI’s function calling
(huggingface.co)
2 points
by
sahil_chaudhary
3y ago
|
0 comments
9.
▲
Instruct-Codegen: open-source instruction following codegen model
(huggingface.co)
4 points
by
sahil_chaudhary
3y ago
|
0 comments
10.
▲
CodeSage: Assistant chatbot to help you code
(codesage.xyz)
1 points
by
sahil_chaudhary
3y ago
|
0 comments
11.
▲
Serverless GPUs without the cold boots
(banana.dev)
3 points
by
sahil_chaudhary
3y ago
|
0 comments
12.
▲
by
sahil_chaudhary
4y ago
Dropping the batch size
13.
▲
by
sahil_chaudhary
4y ago
Yeah the license issue is the only reason why I didn't release the weights yet.
14.
▲
by
sahil_chaudhary
4y ago
I'm training a 65B model right now, also I believe you can use lora-alpaca to train on this data on a much smaller machine.
15.
▲
by
sahil_chaudhary
4y ago
Given how many people are asking for the weights, I'm working to make them available now. So you'll be able to run them locally as well.
16.
▲
by
sahil_chaudhary
4y ago
Yeah the weights did change.
17.
▲
by
sahil_chaudhary
4y ago
All included it costs under 70$ for the 13B model. Training 65B now so will report what that will cost.
18.
▲
by
sahil_chaudhary
4y ago
Interesting, in my experiments alpaca-lora performed worse but definitely possible that isn't the case generally.
19.
▲
CodeAlpaca – Instruction following code generation model
(github.com)
159 points
by
sahil_chaudhary
4y ago
|
47 comments
20.
▲
Show HN: CodeAlpaca – Instruction following model for code
(github.com)
6 points
by
sahil_chaudhary
4y ago
|
0 comments
21.
▲
CodeAlpaca – LLaMA based instruction following code generation model
(github.com)
2 points
by
sahil_chaudhary
4y ago
|
0 comments
22.
▲
CodeAlpaca – Instruction following model to generate code
(code-alpaca-demo.vercel.app)
2 points
by
sahil_chaudhary
4y ago
|
0 comments
23.
▲
I asked GPT-4 to explain CUDA
(sahil-ch.notion.site)
1 points
by
sahil_chaudhary
4y ago
|
0 comments
24.
▲
GPT-4 Explains CUDA
(sahil-ch.notion.site)
3 points
by
sahil_chaudhary
4y ago
|
0 comments
25.
▲
AI Designed Apparel
(prompt-clothing.com)
1 points
by
sahil_chaudhary
4y ago
|
0 comments
26.
▲
by
sahil_chaudhary
4y ago
That was an issue on my end, should be fine now
27.
▲
by
sahil_chaudhary
4y ago
You can find my email on my HN profile
28.
▲
by
sahil_chaudhary
4y ago
Ah yes, this model treats this purely as image captioning. The model isn't open source yet.
29.
▲
by
sahil_chaudhary
4y ago
Nice, does this use clip-interrogator?
30.
▲
by
sahil_chaudhary
4y ago
Glad you enjoy it! It hasn't been trained on non-generated images so is unpredictable when uploading a real photo, specially one with people in it
More ›