HackerTrans
トップ新着トレンドコメント過去質問紹介求人

amasad

no profile record

投稿

Using Tvix Store to Reduce Nix Storage Costs by 90%

blog.replit.com
2 ポイント·投稿者 amasad·昨年·0 コメント

[untitled]

1 ポイント·投稿者 amasad·2 年前·0 コメント

Skip the README, let us install for you

blog.replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

Prioritizing Employee Liquidity at Replit

blog.replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

AI Agent Code Execution

blog.replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

Replit ModelFarm: Zero-Config LLM App Dev Environment

blog.replit.com
2 ポイント·投稿者 amasad·3 年前·0 コメント

Autoscale and Static Deployments

blog.replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

Speeding Up Deployments with Lazy Image Streaming

blog.replit.com
5 ポイント·投稿者 amasad·3 年前·0 コメント

Kaboom.js 3000

kaboomjs.com
140 ポイント·投稿者 amasad·3 年前·13 コメント

Local Code Chatbot Running on 2GB RAM

twitter.com
1 ポイント·投稿者 amasad·3 年前·1 コメント

Ghostwriter Chat: The first conversational AI programmer

blog.replit.com
2 ポイント·投稿者 amasad·3 年前·1 コメント

Startup Replit launches a ChatGPT-like bot for coders

semafor.com
16 ポイント·投稿者 amasad·3 年前·0 コメント

Git in the Shell

blog.replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

Kaboom 3000

3000.kaboomjs.com
2 ポイント·投稿者 amasad·3 年前·0 コメント

Replit Bounties

replit.com
1 ポイント·投稿者 amasad·3 年前·0 コメント

Serverless PostgreSQL on Replit

blog.replit.com
36 ポイント·投稿者 amasad·3 年前·9 コメント

コメント

amasad
·7 か月前·議論
We do let you change it; under your account setting there is a email with an edit button next to it. And it shouldn’t need to match your GitHub email.

Perhaps there is some weirdness if you’ve signed up with GitHub. Feel free to email me and we can take a look: [email protected]
amasad
·3 年前·議論
We're on a budget :) trained on 128 H100-80GB GPUs for a week (200B tokens over 5 epochs, ie 1T tokens).

Tech talk here with timestamp: https://www.youtube.com/live/veShHxQYPzo?si=UlcU9j2kC-C4oWvj...
amasad
·3 年前·議論
Crates is especially challenging -- the build process is very expensive on all resources (cpu, ram, and disk), and packages are very hard (impossible?) to cache. It works super well on the Pro plan.
amasad
·3 年前·議論
It’s communication meant to existing users and it’s a nuanced picture worth elaborating. But if you are not an active user it’s best if you just look at our pricing page for a complete picture: https://replit.com/pricing
amasad
·3 年前·議論
This is for hosting. Development remains free
amasad
·3 年前·議論
Yes, in fact we recently made the free plan more powerful with non-preemptible VMs, premium networking, and package caching for everyone. And working on an AI free tier coming soon. Will likely invest more in the free plan once the abuse from unlimited hosting go away.

Also — the title is slightly misleading. Static hosting is still free, and autoscale (with scale to zero) will effectively cost next to nothing for most users. Especially new programmers get very little traffic that according to our historic data it will cost $0.2 / month.

Finally beginners who cannot afford to pay but learned enough coding can make money on the site. eg, a story earlier today: https://x.com/replit/status/1709577707435274414?s=46
amasad
·3 年前·議論
the blog is slightly outdated — 3000 is out of beta and can be used now. Here is a fun game built with it: https://replit.com/@slmjkdbtl/Bean-Survivor
amasad
·3 年前·議論
Is this the best you can find? not even top 10 bangers.
amasad
·3 年前·議論
Interesting. This seems like a weakness of natural language understanding. If you rephrase your prompt slightly it would get it right. Try:

  // return even numbers that are also more than 10
  const arrayFilter = (array) =>
It would do the right thing. The fine-tuned version gets your prompt right so maybe it benefited from natural language data. Will look more into it.
amasad
·3 年前·議論
Yes we have a robust extension system and some are already building alternatives.
amasad
·3 年前·議論
Broadly finetuning is any post pretraining training. Most of the time it is used in the context of fitting a more narrow task. In our case, it was the same training objective as the pretraining but meant to be more representative of what Replit users like to code. However, we were surprised by how well it boosted overall performance. Best guess: it's a) novel data and b) the model could take even more training!!
amasad
·3 年前·議論
The model is not RLHF'd or instructed. It's an inline autocomplete model so it will get confused if you talk it like you're talking to a person. Altho it is possible to finetune it this way. To get better full function completion try giving it the function definition and a descriptive docstring as a prompt.
amasad
·3 年前·議論
Can't find it now but pretty sure BigCode said somewhere they explicitly looked for it and removed it. Also subjective measure does match up to the benchmark. Our finetuned model performed +50% on HumanEval and then when using it felt at least that much improved.
amasad
·3 年前·議論
1- We trained on languages that are most popular on Replit. Markdown is important because you need some amount of natural language in the data, and it will act as a sort of "natural language label" for code.

2- I like how portable it is being a single small model doing a lot of languages. Single code models are an approach that models like Salesforce/Codegen did that, but I believe we beat (or get very close) to their mono models on benchmarks.
amasad
·3 年前·議論
Hopefully on phones too
amasad
·3 年前·議論
it is 2.7B
amasad
·3 年前·議論
LLMs generally but more so small models will keep going and generate seemingly unrelated things. On the frontend tools like Copilot and Ghostwriter do a lot of things like use stopwords or simply not show completions outside a single block.

As for your prompt, it's following your prompt a little too closely and generating just the function. You can however condition it that this is the start of the program it will do the import, e.g.

   # python function that returns a random integer between min and max
   import
This is in fact a suggestion from OpenAI on best practices for prompting called "leading words" https://help.openai.com/en/articles/6654000-best-practices-f...
amasad
·3 年前·議論
Some links:

- Repo: https://github.com/replit/ReplitLM/tree/main/replit-code-v1-...

- HuggingFace: https://huggingface.co/replit/replit-code-v1-3b

- Demo: https://huggingface.co/spaces/replit/replit-code-v1-3b-demo

- Early benchmark results: https://twitter.com/amasad/status/1651019556423598081

A lot about this project was surprising. We knew it was going to be good, but didn't expect to be this good -- especially surprising was the finetuned performance boost, and the fact that the model is decent at language tasks and reasoning (in some cases much better than much larger general-purpose models).

It feels like there is a lot more to do with this model, and I have a suspicion you can even make a half-decent chatbot (at least one focused on code) by finetuning it on conversation (and/or instruction) datasets.

Will follow up with a more comprehensive technical report and the UL2R version (fill-in-the-middle support).
amasad
·10 年前·議論
Hi Alan,

You've been involved in visual programming environments like GRAIL and Etoys for kids. What do you think of the current state of visual programming for both kids and adults?