Brinens and Things

(part 3)

---

Circa 2022-10-21:
holy shit

apparently there was a bug in Whisper that was causing it to run far more slowly than it had any right to, *especially* on CPU

it's fixed now

(also they've added an option to change how many CPU threads a Whisper instance uses, so if you have 4 threads you can just use 4 and not have to run a pair of Whispers to make use of it all)

I ran a quad-threaded Whisper overnight on 6.5 hours of audio, and nine hours later when I checked on it again *it was already done*


2022-11-19, 10:30 AM:
I haven't run direct head-to-head tests, but it seems as if quad-threaded Whisper is *more* than twice as fast as dual-threaded? It also seems--and *this* makes intuitive sense--that "--condition_on_previous_text=False" runs more slowly (but is more likely to notice when someone starts talking after a long stretch of background noise).

I made the mistake, 35 hours ago, of putting a second one on because it was nearly bedtime and the first one (which was not conditioning on previous text) only had about forty minutes left. The first one's not done, and the second one (conditioned on previous text) has only made it through about 80 minutes.

OTOH, that's making me wonder if my octa-threaded smartphone would be a better cruncher than I'd expected. Perhaps I'll try it at the end of the workweek.


2022-11-19, 11:00 AM:
Ah, you press *ctrl-Z* in a terminal window to pause its active process, and *fg* to resume. Good: it really seemed like there ought to be a way of doing that, but I was having trouble finding it.

I'll test it on Whisper once I've got a Whisper instance running that *wouldn't* lose hours of work if it went wrong. (Not the whole 35 hours--it writes the transcripts to disk at the end of each file, not at the end of the sequence of files--but still.)


2022-11-19, 11:30 AM:
That first one finished a file and moved on to the next. I tried pausing and resuming it. CPU monitor says it's running, but it's not outputting anything to the terminal. It's possible it just hasn't done anything worth outputting yet, though.

A bit of poking at the Termux wiki suggests that Termux isn't very conducive to using Whisper, in any case.

Hmm. It's not the only computer in the house with eight threads. Come to think of it, I wonder if Mom would let me run a password-protected VM inside *her* laptop...


2022-11-19, 11:47 AM:
Oh hey, it's started outputting. I guess I'll try poking Termux to see if it's easier than it looks from the wiki.

liveblog of Termux troubleshooting, click to open (even if you're viewing this post directly and not on a Dreamwidth /read page)

2022-11-19, 12:00 PM:
It can't install because...my version of Python is too *new*? Well, that's not the error message I expected.

Is my August Termux backup an old enough Python version...yes. Let's try reverting to that...


2022-11-19, 12:10 PM:
It still says it can't install torch, but no longer claims that the reason is because of having Python 3.11.


2022-11-19, 12:40 PM:
Turns out that you *can* install torch on Termux, but you have to run pkg install python-torch instead of going through pip. Now the error message is about missing Rust, which the Whisper readme discusses as a common error.

(Another 380 MB of disc space for Rust. Man, this shit is *expensive*. Well, what else was I using my remaining 4.5 GB of internal storage for; as for the external storage, Brother bought me a 512 GB microSD as a birthday present and it's in the mail.)


2022-11-19, 12:55 PM:
"CANNOT LINK EXECUTABLE 'rustc': library 'libLLVM-15.so' not found"

...what if I install libllvm?


2022-11-19, 1:50 PM:
...and aarch64-linux-android-ar...

...no, Termux apt doesn't recognise that one. Hmm. cargo install aarch64-linux-android-ar?

No, not that either. Binutils?


2022-11-19, 2:30 PM:
And once again I have run out of error messages but it's still failing. I wonder what happens if I do something about the warning that wheel isn't installed and it's having to use legacy setup...

2022-11-20, 11:30 AM:
(Meanwhile, on the TV-brain front: suspending the process, understandably enough, doesn't free up RAM. Therefore, pausability does not help.)

more troubleshooting

2022-12-01, 10:30 AM:
Upon closer inspection a few days back, I noticed there *was* another error message.

I got a chance to try the workaround late last night. Half an hour later it was still working at building a wheel for tokenizers, and I went to bed and left it to it.

This morning I checked on it, and...it says everything completed successfully.

Well then. Let's try it out.


2022-12-01, 11:50 AM:
Holy shit, just *unzipping an audio folder* so that Whisper would have something to chew on took like an hour. Not a great sign for how fast this is going to be, though there may have been extenuating circumstances (for one thing, the zip is on the external storage and the unzipped folder is on the internal storage).

Anyway, Material Files has finished that part now, so here we go.

cd /sdcard/2022-09-19 && whisper *.mp3 --model small.en --threads 8 --output_dir /sdcard/2022-09-19-v1


2022-12-01, 12:00 PM:
It...doesn't parse wildcards? Unlike on Linux?

Oh, it's because the unzipping left the folder more nested than I expected, and actually the cd should be /sdcard/2022-09-19/2022-09-19.


2022-12-01, 12:05 PM:
I checked Material Files and internal-storage folder 2022-09-19-v1 has been created, so that's a good sign.

2022-12-02, 12:30 AM:
We're twelve and a half hours in. The first speech in this clip occurs two and a half minutes in. It hasn't outputted any progress to the terminal yet.

I'll let it run overnight mostly out of curiosity, but currently I expect that this device cannot materially contribute to a grid computer. I think tomorrow I'll revert Termux to that August backup again and update from there, freeing up the large quantities of space currently being spent on Whisper and its prerequisites.

The good news is that the phone is not appreciably warm, so I'm not frying my battery.


2022-12-02, 12:30 PM:
We're now at 24.5 hours real-time, and...huh, 11.5 minutes. It did actually manage something.

Still, it looks like my initial hypothesis (before multi-threading) that the phone's effectiveness would be roughly on par with the TV brain was right, and I agree with yesterday-me that it's not worth including this in my available compute.

Neat proof of concept, though.

(part 1)

---

Computers that I physically possess, for which I do not need permission before executing software on them, and on which nobody outside my household stakes a claim:

* My laptop
* Our TV's prosthetic brain (a thirteen-year-old ThinkPad running Linux Lite, with its A/V wired into the TV)
* My smartphone
* Possibly the other two remaining smartphones from my failed experiments with LineageOS, though they may no longer be usable enough: last I checked one of them had a broken USB port and could only be charged by swapping out its battery (which seems unsuitable for extended crunching projects), and I *think* the other one has started bootlooping though I'm not sure if it managed to pull out of that. (...oh hey, it *did* pull out of that. Interesting. I might be able to wring some more compute out of it if I strangle its fucking bloatware first. If it is to be one of my computers (particularly one crunching sensitive data), it will be *mine*, and AT&T gets no say in it whatsoever.)

((even if AT&T doesn't give a shit, it's the principle of the thing))


2022-09-25, 10:50 AM:
Successfully installed Whisper on the TV brain and got it to start crunching 2020-01-22, the first day. It's started outputting and everything! Looks like past-me is discussing electricity prices with her family.

I admit I did not especially understand the instructions I looked up on how to fix the "/home/netflix/.local/bin is not in PATH" errors, but I did successfully infer how to *circumvent* the problem of it not being in PATH:

/home/netflix/.local/bin/./whisper *.mp3 --model small.en --output_dir 2020-01-22-v1

(I realised pretty quickly that I should be labelling the transcript folders "v1", so that when I'm later upgrading them to higher-quality transcripts I'll know which days have been upgraded by how much.)


2022-09-25, 10:30 PM:
Me, last night: I'm pleased that it seems to already have a decent corpus of th-fronters

Whisper: We used to have 12 baddies, now we have a free.

("We used to have twelve bags, now we have three.")

Well, so be it. (And hey, I run into this problem with *people* sometimes too.)


2022-09-25, 10:45 PM:
Checked on the TV brain. Whisper does *run* on a thirteen-year-old laptop, and I appreciate that you're not *completely* locked out of using Whisper if that's all you have, but also it's been crunching for twelve hours and it's made it through about eight minutes. At this rate it's going to take...about four days to get through 2020-01-22, and that day was among the shorter ones.

The smartphones are loosely equal in specs to the TV brain, so I doubt they'd fare much better even if I can figure out an appropriate Termux setup. Definitely doesn't sound feasible to do on my real phone (even considering that I don't touch my phone many if not most days, there would be interruptions multiple times a week). Not sure it's worth trying to get it running on the other one.

...ooh, apparently (for compatibility reasons, I think) Whisper defaults to CPU and you have to explicitly tell it to use a GPU. I'll have to check what happens if I try that on my laptop, though I'll wait for one of the current two to complete rather than trying to throw in a third one. (I have *some* sort of dedicated graphics system, though it wasn't a priority of mine at the time so I don't have a clear sense of what (and, again, six-year-old business laptop, so whatever-it-is probably isn't that good: might be helpful vs CPU-only, though).)


2022-09-26, 8:04 PM:
I didn't leave my laptop running overnight, for...reasons best discussed in a separate post, but (unsurprisingly) the Whispers were undamaged by overnight suspension and one of the initial two has since finished.

I *think*, if I'm making the correct inferences from these error messages and help documents, that my GPU is incompatible because it's *both* old *and* AMD. New AMD (specifically, AMD with ROCm support) is slightly tricky but salvageable; old Nvidia might be salvageable; old AMD doesn't seem to have any salvage options.

Also, Mom says the TV brain is incapable of running Whisper and Netflix at the same time, which means an infeasible number of interruptions on *that* device too. And the AT&T smartphone seems to be locked down hard enough that I can't turn on USB debugging and strangle the bloatware. *God* I hate this phone.

I'll see if I can accumulate any castoff devices for Project House Community Grid, but at this rate I think I'm going to be starved for compute for the next...hmm...127 GB of data...28.8 MB/hour...implies an average of 4.5 hours of new data per day and about 4,400 hours to catch up on...three hours real-time to process one hour of data on my laptop if running two Whispers in parallel...three and a half years.

Well, with a figure like that I'll probably need a new laptop before then anyway. I will definitely be on the lookout for GPU specs next time I'm on the hunt for a primary computer.

In the meantime, I think I'll probably compromise by *not* attempting to catch up on the backlog, and just transcribe going forward for the time being. My poor computer deserves a chance to rest sometimes.

(...did Brother upgrade his gaming desktop's GPU a few months ago? He did some sort of desktop hardware upgrade; I think it might have been that. I wonder if he still has the old one, and what it would take to cobble it into a minimum viable desktop...)

---

(edit: part 3)

(previously on)

---

The Youtube auto-transcriber now gives you the option of seeing the entire transcript at once, rather than having to wait for each line of caption to pop up: click on the three-dot menu (next to "Save"), then click "Open transcript". And yes, that podcast I was looking at earlier has a Youtube mirror now.

The main downside is that it doesn't distinguish between different speakers. That's fine for one-person speeches, but gets a bit confusing with a three-host podcast. There's also still some issues with homophones, obscure proper nouns, and occasionally mistaking obscure words for more common ones that aren't *quite* homophones, but it's pretty easy to figure out what it meant on those.

It was interesting to note that the auto-transcriber *does* know how to spell "nanowrimo", and knows that a string of words occurring after "you can also find us on twitter or instagram at" should be run together into one word.

(inspired by Book Review: The Seven Principles For Making Marriage Work)

It weirds me out when people talk about I-statements and other such speaking techniques.

Is this an autism thing? Like, are allistics actually capable of doing that anywhere close to reliably (assuming they choose to)?

I...basically, I am a Mass Effect character. I can generally control the *gist* of what I am saying (barring the occasional misclick), but for the most part I cannot control the exact words that come out of my mouth: I just have to hope that the phrasing doesn't have connotations or implications that I didn't intend. It's hard enough to make sure I always have the right number of *negations*, and you want me to use *I-statements*??

[cw: amnesia, Maze Runner spoilers]

( Read more... )

---

(This post brought to you by watching the CinemaSins sporkings of the Maze Runner movies. I've been finding that CinemaSins are pretty good for jogging, because they come with subtitles so you don't have to be able to clearly hear what he's saying over the sound of the treadmill. I do wish, though, that he'd consistently subtitle what the characters are saying and not just what he's saying: occasionally he does, but mostly he doesn't. Fairly often he reacts to something and I didn't catch the thing he's reacting to.)