<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: zhinit</title><link>https://news.ycombinator.com/user?id=zhinit</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 16 Jul 2026 22:58:05 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=zhinit" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>I just added one with the "techno" keyword in the keywords section. Its pretty weird lol</p>
]]></description><pubDate>Thu, 16 Jul 2026 21:31:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=48940533</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48940533</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48940533</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>To do this, I think you are right that you would need to 'downgrade' modern recordings to sound old so that you have both sides of the training data covered.<p>This would be a cool project to work on. Ideally you would buy some vintage gear and then run the audio through both, but that would be very expensive. You could may be find some vst emulations though and get decent results.</p>
]]></description><pubDate>Thu, 16 Jul 2026 20:29:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=48939804</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48939804</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48939804</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>Interesting! I had not seen this. On their website they mention diffusion but not the other models so it might not be identical but its definitely similar.</p>
]]></description><pubDate>Thu, 16 Jul 2026 20:23:58 +0000</pubDate><link>https://news.ycombinator.com/item?id=48939741</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48939741</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48939741</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>I sound design a lot of stuff (in fact I made some of the default kicks in the app), but this is just a different tool, and I wanted some practice training and deploying a generative AI model.</p>
]]></description><pubDate>Thu, 16 Jul 2026 20:17:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=48939659</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48939659</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48939659</guid></item><item><title><![CDATA[New comment by zhinit in "$100 AI Music Video: Claude Fable 5 vs. GPT-5.6 Sol"]]></title><description><![CDATA[
<p>Seems like if you build some more scaffolding around it, it wouldn't be bad. I think AI video isn't quite there yet so you probably would want to lean into that. For example you could ask for an animated or cartoon music video so the real shots don't look weird. Also if you gave it some guidance on what a good music video is like it would probably help as well. But yeah idk may be that's not the goal here.</p>
]]></description><pubDate>Thu, 16 Jul 2026 20:14:14 +0000</pubDate><link>https://news.ycombinator.com/item?id=48939624</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48939624</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48939624</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>If you are committed the model should work about the same on any type of one shot sample. The code is public and documented so if you have the snare collection and a macbook you could probably point claude/chatgpt at it and it would be able to train on your laptop.</p>
]]></description><pubDate>Thu, 16 Jul 2026 18:48:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=48938597</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48938597</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48938597</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>Thanks!</p>
]]></description><pubDate>Thu, 16 Jul 2026 18:43:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=48938510</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48938510</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48938510</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>SUNO is pretty close. It still has some weird things going on with high frequency artifacts and phase between left and right channels but if you aren't listening on a good system (like a phone) most people probably wont notice.</p>
]]></description><pubDate>Thu, 16 Jul 2026 18:42:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=48938490</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48938490</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48938490</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>synplant is a great synth!</p>
]]></description><pubDate>Thu, 16 Jul 2026 18:39:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=48938451</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48938451</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48938451</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>The compression is the OTT which stands for Over The Top compression. It was originally a multiband compressor preset in ableton and is now used widely throughout dance music.</p>
]]></description><pubDate>Thu, 16 Jul 2026 17:39:58 +0000</pubDate><link>https://news.ycombinator.com/item?id=48937686</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48937686</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48937686</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>the spectrograms are 128x173 (128 mel frequency bins by 173 time frames)
the encoder is downsampling 4 stages of stride 2 convolutions so it halves dimensions 4 times<p>0: 128 x 173<p>1: 64 x 87<p>2: 32 x 44<p>3: 16 x 22<p>4: 8 x 11<p>Then i used 4 separate channels.<p>This was somewhat arbitrary due to the local training constraint. This would be a hyper parameter worth tuning if I had time to dig into this more.<p>I trained this a few month ago and don't remember exactly what I tried before I arrived here, but I only ran the whole process 2 or 3 times because of how long it took to train.
Hope this answers your question!</p>
]]></description><pubDate>Thu, 16 Jul 2026 17:37:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=48937663</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48937663</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48937663</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>That was an oversight. The tower is more than 10 years old. I updated the article to be exact now</p>
]]></description><pubDate>Thu, 16 Jul 2026 17:17:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=48937372</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48937372</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48937372</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>For sure.<p>If you are curious I used a NVIDIA GeForce GTX 1660 SUPER
So to be exact, it came out 7 years ago (I upgraded at some point on this desktop a long time ago and didn't remember the exact year) (I updated the article to reflect this now)<p>This cost $230 new and you can get one now for $100 which I don't think is too out of reach.</p>
]]></description><pubDate>Thu, 16 Jul 2026 16:50:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=48936987</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48936987</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48936987</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>For sure! I just added a couple</p>
]]></description><pubDate>Thu, 16 Jul 2026 16:37:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=48936831</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48936831</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48936831</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>Good point! I just added a couple at the end of the intro</p>
]]></description><pubDate>Thu, 16 Jul 2026 16:37:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=48936825</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48936825</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48936825</guid></item><item><title><![CDATA[New comment by zhinit in "How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM"]]></title><description><![CDATA[
<p>I wouldn't exactly say it's trying to solve a problem. It's to explore and see what happens which is what music is all about. It's also a unique niche model I haven't seen before.</p>
]]></description><pubDate>Thu, 16 Jul 2026 16:20:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48936604</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48936604</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48936604</guid></item><item><title><![CDATA[How to Train a Gen AI Kick Drum Model on Your Old Linux Desktop with 6GB VRAM]]></title><description><![CDATA[
<p>Article URL: <a href="https://www.zhinit.dev/blog/training-a-kick-drum-diffusion-model">https://www.zhinit.dev/blog/training-a-kick-drum-diffusion-model</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48935687">https://news.ycombinator.com/item?id=48935687</a></p>
<p>Points: 82</p>
<p># Comments: 52</p>
]]></description><pubDate>Thu, 16 Jul 2026 15:13:00 +0000</pubDate><link>https://www.zhinit.dev/blog/training-a-kick-drum-diffusion-model</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48935687</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48935687</guid></item><item><title><![CDATA[Detecting Prompt Injection Attacks on Purpose-Specific LLM Agents]]></title><description><![CDATA[
<p>Article URL: <a href="https://arxiv.org/abs/2607.12624">https://arxiv.org/abs/2607.12624</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48921816">https://news.ycombinator.com/item?id=48921816</a></p>
<p>Points: 3</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 15 Jul 2026 14:56:23 +0000</pubDate><link>https://arxiv.org/abs/2607.12624</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48921816</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48921816</guid></item><item><title><![CDATA[Prismata: Confining cross-site prompt injection in web agents]]></title><description><![CDATA[
<p>Article URL: <a href="https://arxiv.org/abs/2607.08147">https://arxiv.org/abs/2607.08147</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=48865238">https://news.ycombinator.com/item?id=48865238</a></p>
<p>Points: 14</p>
<p># Comments: 0</p>
]]></description><pubDate>Fri, 10 Jul 2026 21:03:38 +0000</pubDate><link>https://arxiv.org/abs/2607.08147</link><dc:creator>zhinit</dc:creator><comments>https://news.ycombinator.com/item?id=48865238</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48865238</guid></item></channel></rss>