Show HN: Shoehorn – Quantize any model down to run on your machine • Hacker News Board

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#11

Post by kelvo_ran » Thu, Aug 20, 2026, 1:41 AM UTC

[dead]

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#12

Post by jedbrooke » Fri, Aug 21, 2026, 7:02 PM UTC

I gotta laugh at some of the models it suggests, for example:

> AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF

you’re telling me you managed to fit Fable 5 into just 4B?

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#13

Post by chompychop » Fri, Aug 21, 2026, 7:28 PM UTC

jedbrooke wrote: post #12

I gotta laugh at some of the models it suggests, for example: > AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF you’re telling me you managed to fit Fable 5 into just 4B?

[flagged]

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#14

Post by unrented7977 » Fri, Aug 21, 2026, 7:48 PM UTC

chompychop wrote: post #13

jedbrooke wrote: post #12

I gotta laugh at some of the models it suggests, for example: > AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF you’re telling me you managed to fit Fable 5 into just 4B?

[flagged]

Don't make fun of people you think are ignorant, it's a pretty shitty look

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#15

Post by chompychop » Fri, Aug 21, 2026, 8:04 PM UTC

unrented7977 wrote: post #14

chompychop wrote: post #13

Earlier quoted context omitted.

[flagged]

Don't make fun of people you think are ignorant, it's a pretty shitty look

[flagged]

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#16

Post by sscarduzio » Fri, Aug 21, 2026, 8:27 PM UTC

The project name is perfect!

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#17

Post by metalliqaz » Fri, Aug 21, 2026, 8:54 PM UTC

chompychop wrote: post #13

jedbrooke wrote: post #12

I gotta laugh at some of the models it suggests, for example: > AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF you’re telling me you managed to fit Fable 5 into just 4B?

[flagged]

Well to be fair here... the title of this post doesn't mention fine tuning, it mentions quantization.

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#18

Post by metalliqaz » Fri, Aug 21, 2026, 8:54 PM UTC

kennywinker wrote: post #8

rhgraysonii wrote: post #6

Earlier quoted context omitted.

Yes that is exactly what this does.

Could you explain what happens when you try to shoehorn a 2.4T parameter model into a 24gb m4 mac?

extreme divergence would be my guess

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#19

Post by puttycat » Fri, Aug 21, 2026, 9:14 PM UTC

This is really impressive. Can you say a bit about the underlying process? I'm guessing this is post-training qantization? Isn't PTQ also resource-intensive? (Ie might not work on any machine)

Re: Show HN: Shoehorn – Quantize any model down to run on your machine

#20

Post by mannyv » Sat, Aug 22, 2026, 4:31 AM UTC

jedbrooke wrote: post #12

I gotta laugh at some of the models it suggests, for example: > AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF you’re telling me you managed to fit Fable 5 into just 4B?

Fyi that model name to me reads

Qwen3 4b params distilled/trained with fable 5