GithubHelp home page GithubHelp logo

missing lines about whisper-standalone-win HOT 6 CLOSED

 avatar commented on June 12, 2024
missing lines

from whisper-standalone-win.

Comments (6)

Purfview avatar Purfview commented on June 12, 2024

Upload the audio to a normal download site.

from whisper-standalone-win.

 avatar commented on June 12, 2024

https://we.tl/t-hvSN7uy7JH

from whisper-standalone-win.

Purfview avatar Purfview commented on June 12, 2024

Transcribes for me with whisper-faster.exe test.mka -m large-v2 -l en -v true -ct float32 -bs 1:

[02:14.040 --> 02:18.740]  But before we jump into the overclocking, let's first have a look at the hardware and the
[02:18.740 --> 02:23.640]  benchmarks that we'll be using in this guide. The system we're overclocking today consists of
Processing segment at 02:23.640
[02:23.640 --> 02:29.740]  the following hardware. The specific graphics card we're using to overclock the RTX 3050 GPU
[02:29.740 --> 02:39.460]  is the EVGA GeForce RTX 3050 XC Gaming. Of course, unfortunately, EVGA has recently exited the GPU
[02:39.460 --> 02:46.340]  business. The card has a dual fan thermal solution and an out of the box listed GPU boost frequency
Processing segment at 02:46.340
[02:46.340 --> 02:56.640]  of 1,845MHz, slightly up from NVIDIA's 1,777MHz specification. However, it appears as if EVGA
[02:56.640 --> 03:03.700]  overclocked the GPU even further as we get up to 2,055MHz when we power up the card. We'll get back
[03:03.700 --> 03:09.580]  to that later in this video. We use Windows 11 and the following benchmark applications to measure
[03:09.580 --> 03:14.660]  performance and ensure system stability. I'm introducing a couple of new applications in this
Processing segment at 03:14.660

Post what it shows for you in console at those times. [with bfloat16 and verbose enabled]

from whisper-standalone-win.

 avatar commented on June 12, 2024

with .\whisper-faster.exe --sentence --model=large-v2 -v true --compute_type=bfloat16 --ff_rnndn_sh file.opus

[01:49.960 --> 02:13.360]  In this video we cover 3 overclocking strategies. First we manually increase the GPU and memory frequency using the EVGA Precision X1 software tool. Second, we slap on a full cover waterblock and try to overcome the power limitations by shunt modding. Third, we use the Elmo Labs EVC2 to work around other overclocking limitations.
Processing segment at 02:13.360
[02:13.940 --> 02:21.020]  But before we jump into the overclocking, let's first have a look at the hardware and the benchmarks that we'll be using in this guide.
Processing segment at 02:43.360
[03:04.380 --> 03:11.860]  We use Windows 11 and the following benchmark applications to measure performance and ensure system stability.
Processing segment at 03:11.860

from whisper-standalone-win.

 avatar commented on June 12, 2024

problam was with bfloat16, float16 works gud

from whisper-standalone-win.

Purfview avatar Purfview commented on June 12, 2024

It just doesn't transcribe those lines with bfloat16, no idea why. That's a question for faster-whisper devs.

from whisper-standalone-win.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.