Nice. But, Oh, Uh, please look at this URL: https://huggingface.co/api/spaces/Ark-kun/tangent-1/runtime. Since it says Flagged as abusive, that means the Space has been deliberately paused by Hugging Face-side systems, not just accidentally broken by the app. So I think repeatedly creating similar duplicates is risky. Details:

The important part is the runtime API, not the 503…

Hmm… It looks like you may already be past the “which model should I choose?” stage and into the next stage:

You have already done one of the hardest beginner steps: you have a local stack running at all.

Docker Model Runner + Open WebUI + local models + Knowledge collections is not a strange direction. Docker has official docs for using Open WebUI with Docker Model Runner, and Open…

Since whether a model is useful depends heavily on what you want to accomplish, I don’t think there is a simple universal “top 5” I can give you. But I can outline some useful ways to search and evaluate models under that assumption:

Short version

I would not treat this as “find the five best uncensored models.”

I would turn it into a branching process:

use case
→ hardware…

If you want to see why there is such a high refusal rate on goth girl or boho girl types in undelivered images - see what you get in cute minifig form.

Input (click for more details)

For now, here are some general things that might be useful for Paper 1:

The staged roadmap makes the project much easier to reason about. For Paper 1 , I would try to make the first artifact a reusable measurement map , not a proof of UCTF and not a representation design yet.

My direct suggestion would be:

Separate the measurement paths. Use clean aligned data for sanity…

I looked into some practical know-how for using OSS AI on a self-hosted server while assuming the network may be unstable:

Since your workstation itself is not the bottleneck, I would frame this as a travel reliability / offline packaging problem, not mainly as a “which GPU/model is enough?” problem.

The short version:

**Using a home machine as a remote OSS AI box is a real and…

Oh. If I know the hardware specs, I can narrow the candidates down quite a bit. As for unstable networks while traveling… hmm… what to do about that​:

Direct answer

Given that workstation, I would treat this as two separate problems:

1. On the workstation: choose realistic local baselines for a 12GB VRAM GPU, then test them with a small eval set.
2. While traveling: build…

There are a few fully multimodal / omni-style large models, but if the more general goal is “I want my OSS local chat/RAG setup to call T2I/T2V”, I would usually make the image/video models separate local server processes and connect them as a pipeline. The execution cost, debugging cost, and replacement cost are usually lower that way. Existing frameworks already cover a lot of this:

##…

ia0:

All crates can cause undefined behavior, even if they use #![forbid(unsafe_code)].

Let's consider a crate author that doesn't want to deal with undefined behavior. They write a crate foo which uses only safe code and forbids unsafe code. Another crate author writes a crate bar which depends on foo. They write unsafe code which relies on the safety guarantees of foo (which is…

I think that’s probably the safer direction. If I were organizing that route, I’d frame it like this:

Short read

This route becomes easier to evaluate if each branch gets its own narrow label.

For the speech branch:

tokenizer-free speech on the fixed HSL substrate

For the BPE text-generation branch:

**embedding-table-free text generation over BPE-segmented HSL…

Well, If I can assume your technical stack, the explanation can be fairly dense:

Direct answer

I would not start by looking for “the best model.” I would first split the problem into layers:

• local runtime
• model format and quantization
• chat template
• a small eval set
• RAG / retrieval
• fine-tuning / adapters
• model choice
• tool use / agents
• offline and…

Codex Desktop macOS GPU spike when focused

Summary

On macOS, the Codex Desktop app appears to trigger heavy sustained GPU/compositor activity whenever the Codex window is focused/frontmost. Switching focus to Chrome drops the GPU monitor to near idle immediately.

The effect reproduces with a small A/B script that focuses Chrome, samples powermetrics, then focuses Codex and samples…

Hmm… with only a 503 error, it is hard to narrow this down. If you can see a concrete error in the build or container logs, there may be cases where this can be fixed from the user/repo side. But if no useful logs appear, issues like this often need Hugging Face support to check the Space from their side, e.g. viawebsite@huggingface.co:

I would first separate this into two different…

Chat handoff is always a headache…

I like this framing. I would not read GCA as only “more memory” or “a better summary.” The interesting part, to me, is that it treats handoff itself as a lifecycle problem:

trigger → write → persist → validate → read back → bootstrap successor

That seems useful for long-running agents. The difficult question may not be only _what the final…

Terminology for unsafe (expand if unfamiliar) (click for more details) Terminology for contracts (expand if unfamiliar) (click for more details)

When designing a public API, the logic contract can be freely chosen but the safety contract must follow these constraints:

• The safe part of public APIs must not have safety requirements. Public APIs should avoid having safety requirements.…

Hmm… For now, I’d organize it roughly like this:

I would not start by treating UCTF as either “already solved” or “obviously impossible.” I would first map it to a more testable family of ideas:

multilingual semantic bottlenecks / neural interlingua / language-agnostic sentence representations / concept-space modeling / discrete universal codes.

My direct answer would be:

• I do…

Hmm… there are still a lot of unknowns here, so I’ll just leave some rough thoughts:

My quick feedback would be: the credit idea is interesting, but I would treat trust, validation, and curation as the core design problem.

Credits can help people exchange value, but they do not by themselves answer questions like:

• where did this data come from?
• can it be redistributed or…

Hmm… maybe something like this for now?:

I do not read this as a one-step “accept or reject” result. I would mostly treat it as a useful measurement map, and then separate a few layers so that future readers can tell which part is already measured, which part is a proxy-labeling choice, which part is a probe result, and which part would need later causal testing.

Directly answering your…

Hmm… I looked into it a bit on Colab:

Short read

My read is that the STT side is already a fairly clean path-specific lens result, while the TTS side should probably be treated as a free-run acoustic-path debugging problem , not first as an HSL text-input failure.

For STT, the important result is not “HSL alone solves speech.” It is more specific:

fixed HSL substrate + a…

Hi. Short version: I think this is probably a bug , and probably not something caused by you or your Space:

The most useful way to frame this is probably:

Direct/canonical paths work, but the browser-facing Gradio Web Component / Space identity / host / config-resolution path fails.

So I would not start by changing your Python code, model code, or Gradio layout. Your checks…

Surreal scene of multiple birds, extremely abstract fractured Cubism drawing, loose rapid hand-drawn lines...etc (click for more details)

_seems these AIs can’t rewrite or see anything without it being surreal_

Angry Birds Redux

Prompt (click for more details)

jeffvpace:

> Cooking for Sunday

Vocabulary lessons (click for more details)

_j:

'cause nobody’s sharing

Sorry, the one that introduced the concept of sleep has to be an evil spirit.

Anyway, Good Morning!

Sunday's Idea (click for more details)

Hmm… I don’t have direct experience with this exact setup, but I think this is one of those cases where the right thing to do changes a lot depending on what you ultimately want to achieve :

Short version

I would not start by treating “dissimilar language ablation” as one single operation.

Depending on the actual goal, this could be:

Goal | First thing I’d try | What it does |…

Impossible: not having perspective

This is the most insistent prompt, then followed by single-token items

Knolling photograph

Create image in the style of knolling photography. Items are carefully laid out and photographed. The view is from directly overhead. The camera is looking directly down in rectilinear proportion. The image is taken of a real physical space. Every item is in…

Impossible: not having perspective

This is the most insistent prompt, then followed by single-token items

Knolling photograph

Create image in the style of knolling photography. Items are carefully laid out and photographed. The view is from directly overhead. The camera is looking directly down in rectilinear proportion. The image is taken of a real physical space. Every item is in…

Yeah. That direction is probably the right one:

I think the useful part of your reframing is exactly this: keeping the current claim behavioral for now , while making the stronger mechanistic claims testable rather than asserted.

I cannot promise that I can help at a high technical level. I do not have much compute available, and I do not want to overstate my role. At most, I may…

Yeah. That direction is probably the right one:

I think the useful part of your reframing is exactly this: keeping the current claim behavioral for now , while making the stronger mechanistic claims testable rather than asserted.

I cannot promise that I can help at a high technical level. I do not have much compute available, and I do not want to overstate my role. At most, I may…

Yeah. That direction is probably the right one:

I think the useful part of your reframing is exactly this: keeping the current claim behavioral for now , while making the stronger mechanistic claims testable rather than asserted.

I cannot promise that I can help at a high technical level. I do not have much compute available, and I do not want to overstate my role. At most, I may…