<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>NOT Dexter's Lab - Latest Comments</title><link>http://a2p.disqus.com/</link><description></description><atom:link href="https://a2p.disqus.com/comments.rss" rel="self"></atom:link><language>en</language><lastBuildDate>Sun, 23 Feb 2025 13:10:31 -0000</lastBuildDate><item><title>Re: Pretrained Vision Mamba: a minimal example</title><link>https://www.a2p.it/machine-learning/pretrained-vision-mamba-minimal-example/#comment-6657678527</link><description>&lt;p&gt;Hi Alessio Placitelli,&lt;/p&gt;&lt;p&gt;I am getting this error, do you know where is the issue?&lt;/p&gt;&lt;p&gt;---------------------------------------------------------------------------&lt;br&gt;TypeError                                 Traceback (most recent call last)&lt;br&gt;/mnt/data1/taha/test.ipynb Cell 6 line 1&lt;br&gt;----&amp;gt; 1 model = VisionMamba(&lt;br&gt;      2     patch_size=16,&lt;br&gt;      3     stride=8,&lt;br&gt;      4     embed_dim=384,&lt;br&gt;      5     depth=24,&lt;br&gt;      6     rms_norm=True,&lt;br&gt;      7     residual_in_fp32=True,&lt;br&gt;      8     fused_add_norm=True,&lt;br&gt;      9     final_pool_type='mean',&lt;br&gt;     10     if_abs_pos_embed=True,&lt;br&gt;     11     if_rope=False,&lt;br&gt;     12     if_rope_residual=False,&lt;br&gt;     13     bimamba_type="v2",&lt;br&gt;     14     if_cls_token=True,&lt;br&gt;     15     if_devide_out=True,&lt;br&gt;     16     use_middle_cls_token=True,&lt;br&gt;     17     num_classes=1000,&lt;br&gt;     18     drop_rate=0.0,&lt;br&gt;     19     drop_path_rate=0.1,&lt;br&gt;     20     drop_block_rate=None,&lt;br&gt;     21     img_size=224,&lt;br&gt;     22 )&lt;br&gt;...&lt;br&gt;    175     residual_in_fp32=residual_in_fp32,&lt;br&gt;    176 )&lt;br&gt;    177 block.layer_idx = layer_idx&lt;/p&gt;&lt;p&gt;TypeError: the first argument must be callable&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Taha</dc:creator><pubDate>Sun, 23 Feb 2025 13:10:31 -0000</pubDate></item><item><title>Re: Pretrained Vision Mamba: a minimal example</title><link>https://www.a2p.it/machine-learning/pretrained-vision-mamba-minimal-example/#comment-6576383954</link><description>&lt;p&gt;Hey, as mentioned in the article it's important to use CUDA 11.8 :)&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Wed, 23 Oct 2024 02:46:00 -0000</pubDate></item><item><title>Re: Pretrained Vision Mamba: a minimal example</title><link>https://www.a2p.it/machine-learning/pretrained-vision-mamba-minimal-example/#comment-6576353816</link><description>&lt;p&gt;RuntimeError:&lt;br&gt;    The detected CUDA version (12.3) mismatches the version that was used to compile&lt;br&gt;    PyTorch (11.8). Please make sure to use the same CUDA versions.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">sohag hossain</dc:creator><pubDate>Wed, 23 Oct 2024 00:54:50 -0000</pubDate></item><item><title>Re: Pretrained Vision Mamba: a minimal example</title><link>https://www.a2p.it/machine-learning/pretrained-vision-mamba-minimal-example/#comment-6421799903</link><description>&lt;p&gt;Hey, I'm glad you liked the article! Unfortunately I wasn't able to make this work on CUDA 12, as the kernels used for conv1d required some changes. I tried to use the most recent version of conv1d and mamba-ssm (which support CUDA 12), but then the model would not work.&lt;/p&gt;&lt;p&gt;I ended up using CUDA 11.8. I wrote a short article to switch between CUDA versions on Ubuntu: &lt;a href="https://www.a2p.it/machine-learning/different-cuda-versions-in-ubuntu/" rel="nofollow noopener" target="_blank" title="https://www.a2p.it/machine-learning/different-cuda-versions-in-ubuntu/"&gt;https://www.a2p.it/machine-...&lt;/a&gt;&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Tue, 26 Mar 2024 05:00:03 -0000</pubDate></item><item><title>Re: Pretrained Vision Mamba: a minimal example</title><link>https://www.a2p.it/machine-learning/pretrained-vision-mamba-minimal-example/#comment-6421763838</link><description>&lt;p&gt;hello Alessio, thank you for the comprehensive explanation. I appreciate.&lt;br&gt;I'm just wondering. Is impossible run the Vision Mamba on CUDA 12.4? Since I already install the CUDA 12 in my system, is not easy to shift to CUDA 11.8.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">姜华为</dc:creator><pubDate>Tue, 26 Mar 2024 03:31:38 -0000</pubDate></item><item><title>Re: 
      CoquiSTT + Signal = Love (death to voice messages)
    </title><link>https://www.a2p.it/tech-stuff/coquistt-signal-love-death-to-voice-messages/#comment-5910686771</link><description>&lt;p&gt;Thank you!&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Fri, 08 Jul 2022 04:26:08 -0000</pubDate></item><item><title>Re: 
      CoquiSTT + Signal = Love (death to voice messages)
    </title><link>https://www.a2p.it/tech-stuff/coquistt-signal-love-death-to-voice-messages/#comment-5910344368</link><description>&lt;p&gt;Such a slick integration!&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Josh Meyer</dc:creator><pubDate>Thu, 07 Jul 2022 18:02:12 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-4614669186</link><description>&lt;p&gt;HELLO, Thanks for the tutorial.&lt;br&gt;I don't see my .a file anymore in the path i have defined.&lt;br&gt;Installed cygwin, putty and tools as mentioned in the tutorials. And even did created the .bat file as mentioned . I am not understanding where i need to mention which file to compile and what to do at linking and where the .a file needs to be . I just mentioned this in " post build steps ": C:/Users/Mari/Downloads/eclipse-cpp-helios-SR2-win32-x86_64/eclipse/deploy_debug.bat ${PWD}${ProjName}.a "/home/pi/projects/${ProjName}.a"  Please let me know what error it can be .Thanks in advance.&lt;br&gt; &lt;a href="https://uploads.disquscdn.com/images/ac4f9dabfb25776057dbfa16013936ea699f18003877b853a012ea351354088f.png" rel="nofollow noopener" target="_blank" title="https://uploads.disquscdn.com/images/ac4f9dabfb25776057dbfa16013936ea699f18003877b853a012ea351354088f.png"&gt;https://uploads.disquscdn.c...&lt;/a&gt;&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">SUMAN MAHENDRA KUMAR</dc:creator><pubDate>Fri, 13 Sep 2019 10:46:01 -0000</pubDate></item><item><title>Re: Getting Firefox data faster: the shutdown pingsender</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/firefox-data-faster-shutdown-pingsender/#comment-4118825576</link><description>&lt;p&gt;This component doesn't have anything to do with updates, so you should be safe with respect to that. The pingsender is used to send the Telemetry that Firefox already sends while running, when it is shut down. I personally think that if an attacker has local access to the system, with high enough privileges to overwrite pingsender.exe, I bet there's nothing that would prevent him for overwriting any other executable or dropping new ones on the filesystem.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Fri, 28 Sep 2018 09:08:23 -0000</pubDate></item><item><title>Re: Getting Firefox data faster: the shutdown pingsender</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/firefox-data-faster-shutdown-pingsender/#comment-4111807551</link><description>&lt;p&gt;Is this program necessary?  Bitdefender threw up all kinds of warnings about it being changed and I blocked the changes.  If I need to use it, how do I go back in and allow for the updates?  Enquiring Minds Want To Know!  &lt;br&gt;After reading the note posted by attitude_check, I am of the mind to say "screw it".&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">UTexasex</dc:creator><pubDate>Mon, 24 Sep 2018 12:45:26 -0000</pubDate></item><item><title>Re: Add-on recommendations for Firefox users: a prototype recommender system leveraging existing data sources</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/add-on-recommendations-for-firefox-users-a-prototype-recommender-system-leveraging-existing-data-sources/#comment-3653940469</link><description>&lt;p&gt;I agree with you, I think that exploring these other features totally makes sense. Measuring the correlation between add-ons and performance metrics is something that can be done (and its being done in other projects). We will have to verify that adding these new features to TAAR would produce a benefit that is greater than the added computational complexity. We're currently brainstorming ideas like that in bug 1286224.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Sat, 09 Dec 2017 04:15:55 -0000</pubDate></item><item><title>Re: Add-on recommendations for Firefox users: a prototype recommender system leveraging existing data sources</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/add-on-recommendations-for-firefox-users-a-prototype-recommender-system-leveraging-existing-data-sources/#comment-3652883609</link><description>&lt;p&gt;I think some of the most useful features that should give recommendations a boost, aside from popularity/ratings, are:&lt;br&gt;* few permissions required (this info currently isn't visible on AMO, AFAIK)&lt;br&gt;* average size of performance hit after users enable the extension (is this available from telemetry somehow?)&lt;br&gt;* produced by known/verified/trusted user&lt;br&gt;* source code publicly available / issue board publicly available&lt;br&gt;* evidence of basic security / privacy-conscious practices in the code, if a metric can be determined for that&lt;/p&gt;&lt;p&gt;Some of these things are really hard to evaluate before installing the extension or even after.  I'd be great to measure them and use them to boost / suppress recommendations.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">AMO user</dc:creator><pubDate>Fri, 08 Dec 2017 12:08:48 -0000</pubDate></item><item><title>Re: Getting Firefox data faster: the shutdown pingsender</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/firefox-data-faster-shutdown-pingsender/#comment-3506442940</link><description>&lt;p&gt;Hey, thank you for reaching out on this and for sharing your concerns! It would be great if you could file a bug about that using Bugzilla (&lt;a href="http://bugzilla.mozilla.org" rel="nofollow noopener" target="_blank" title="bugzilla.mozilla.org"&gt;bugzilla.mozilla.org&lt;/a&gt;): Core-&amp;gt;Build Config seems like a reasonable component. This way we could spawn the proper discussion around this matter and make the security folks address your concerns!&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Thu, 07 Sep 2017 10:55:43 -0000</pubDate></item><item><title>Re: Getting Firefox data faster: the shutdown pingsender</title><link>https://www.a2p.it/wordpress/tech-stuff/mozilla/firefox-data-faster-shutdown-pingsender/#comment-3497824312</link><description>&lt;p&gt;I am concerned this could be a path for future virus and worm communication.  If pingsender.exe is overwritten, it could be used to communicate with anything or anyone.  It is presently only signed with a SHA-1, which we know is getting very vulnerable.  Recommend at least upgrading the signature to a SHA-256, to make an undetected hostile file overwrite more difficult to achieve.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">attitude_check</dc:creator><pubDate>Fri, 01 Sep 2017 14:20:00 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2576073791</link><description>&lt;p&gt;I'm not sure how to help with that, but I saw plenty of discussions here: &lt;a href="https://www.raspberrypi.org/forums/viewtopic.php?f=33&amp;amp;t=50419" rel="nofollow noopener" target="_blank" title="https://www.raspberrypi.org/forums/viewtopic.php?f=33&amp;amp;t=50419"&gt;https://www.raspberrypi.org...&lt;/a&gt;&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Fri, 18 Mar 2016 03:05:50 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2575254523</link><description>&lt;p&gt;I tried downloading the libpython2.6.dll file from a strange site, but I think it was solved. Now when I try to debug, I am getting this error:&lt;/p&gt;&lt;p&gt;Cannot access memory at address 0x0&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Thu, 17 Mar 2016 15:46:21 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2574849014</link><description>&lt;p&gt;The cygncurses-10.dll error was fixed by installing libncurses10 package from the installation setup, in case someone gets this error.&lt;/p&gt;&lt;p&gt;Still getting the libpython2.6.dll error.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Thu, 17 Mar 2016 12:11:51 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2574786135</link><description>&lt;p&gt;I rain the Cygwin installer again and installed all the available core utils, but still getting the same error. I also get a missing libpython2.6.dll error.&lt;/p&gt;&lt;p&gt;I read that 2.6 version is not available anymore. Do you know any way of getting through this?&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Thu, 17 Mar 2016 11:36:40 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2574731404</link><description>&lt;p&gt;Did you install the required dependencies from CygWin? It looks like core utils are missing.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Thu, 17 Mar 2016 11:03:59 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2574727882</link><description>&lt;p&gt;I tried to run it from the command line but I had this error:&lt;/p&gt;&lt;p&gt;This application has failed to start because cygncurses-10.dll was not found. Re-installing the application may fix this problem.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Thu, 17 Mar 2016 11:01:45 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2574317079</link><description>&lt;p&gt;Can you manually run &lt;br&gt;C:\cygwin\opt\cross\x-tools\arm-unknown-linux-gnueabi\bin\arm-unknown-linux-gnueabi-gdb.exe&lt;br&gt; --version from command line?&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Thu, 17 Mar 2016 04:56:55 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2573571924</link><description>&lt;p&gt;I could solve the issue. The problem was because I was putting the "s in the address. I now have this other issue:&lt;/p&gt;&lt;p&gt;Could not determine GDB version using command:&lt;br&gt;C:\cygwin\opt\cross\x-tools\arm-unknown-linux-gnueabi\bin\arm-unknown-linux-gnueabi-gdb.exe --version&lt;/p&gt;&lt;p&gt;Any help would be appreciated.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Wed, 16 Mar 2016 17:53:17 -0000</pubDate></item><item><title>Re: Remote cross debugging: Windows to Raspberry-Pi</title><link>https://www.a2p.it/wordpress/tech-stuff/development/remote-debugging-raspberrypi/#comment-2572955721</link><description>&lt;p&gt;Hi! Thanks for this post. I am having trouble with step 6. The error I get is:&lt;/p&gt;&lt;p&gt;C:/eclipse/deploy_debug.bat HelloCygwin.a “/home/pi/projects/HelloCygwin.a”&lt;/p&gt;&lt;p&gt;pscp: unable to open ?/home/pi/projects/HelloCygwin.a?: no such file or directory&lt;/p&gt;&lt;p&gt;Any ideas? Thanks!&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Oscar Pauly</dc:creator><pubDate>Wed, 16 Mar 2016 12:25:34 -0000</pubDate></item><item><title>Re: Raspberry Pi: tool chain per cross-compilazione su Windows Cygwin</title><link>https://www.a2p.it/wordpress/tecnologia/sviluppo/raspberry-pi-tool-chain-per-cross-compilazione-su-windows-cygwin/#comment-2213240818</link><description>&lt;p&gt;Ciao! Scusa il ritardo :-)&lt;/p&gt;&lt;p&gt;Python 2.6 e Cygwin 32 bit&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Alessio Placitelli</dc:creator><pubDate>Tue, 25 Aug 2015 01:12:36 -0000</pubDate></item><item><title>Re: Raspberry Pi: tool chain per cross-compilazione su Windows Cygwin</title><link>https://www.a2p.it/wordpress/tecnologia/sviluppo/raspberry-pi-tool-chain-per-cross-compilazione-su-windows-cygwin/#comment-2201853155</link><description>&lt;p&gt;Ciao!&lt;br&gt;abbi pazienza:&lt;br&gt;che versione hai usato del cygwin? 32 o 64bit?&lt;br&gt;del python 2.6 o 2.7?&lt;br&gt;Grazie mille&lt;br&gt;Lucio.&lt;/p&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Lucio Alessandroni</dc:creator><pubDate>Tue, 18 Aug 2015 20:22:50 -0000</pubDate></item></channel></rss>