* . *
  • About
  • Advertise
  • Privacy & Policy
  • Contact
Monday, September 15, 2025
Earth-News
  • Home
  • Business
  • Entertainment
    Fall’s Hottest Ski Films Are Going on Tour—Here’s the List – yahoo.com

    Experience the Thrill: Fall’s Hottest Ski Films Hit the Road-Don’t Miss the Tour!

    Ryan Reynolds reveals he called a journalist who said mean things about John Candy – yahoo.com

    Ryan Reynolds Reveals the Moment He Stood Up to a Journalist Who Insulted John Candy

    Entertainment Community Fund Launches Program Supporting Entrepreneurs – Playbill

    Entertainment Community Fund Unveils Exciting New Program to Empower Entrepreneurs

    Behind the turntables: DJ Johnny Kage’s story of perseverance – yahoo.com

    Behind the Turntables: DJ Johnny Kage’s Inspiring Journey of Perseverance

    The other WWE star James Gunn wanted for Peacemaker instead of John Cena – yahoo.com

    The WWE Star James Gunn Originally Wanted for Peacemaker Instead of John Cena

    Quinta Brunson, John Stamos Join Entertainment and Technology Summit – Variety

    Quinta Brunson and John Stamos to Headline Thrilling Entertainment and Technology Summit

  • General
  • Health
  • News

    Cracking the Code: Why China’s Economic Challenges Aren’t Shaking Markets, Unlike America’s” – Bloomberg

    Trump’s Narrow Window to Spread the Truth About Harris

    Trump’s Narrow Window to Spread the Truth About Harris

    Israel-Gaza war live updates: Hamas leader Ismail Haniyeh assassinated in Iran, group says

    Israel-Gaza war live updates: Hamas leader Ismail Haniyeh assassinated in Iran, group says

    PAP Boss to Niger Delta Youths, Stay Away from the Protest

    PAP Boss to Niger Delta Youths, Stay Away from the Protest

    Court Restricts Protests In Lagos To Freedom, Peace Park

    Court Restricts Protests In Lagos To Freedom, Peace Park

    Fans React to Jazz Jennings’ Inspiring Weight Loss Journey

    Fans React to Jazz Jennings’ Inspiring Weight Loss Journey

    Trending Tags

    • Trump Inauguration
    • United Stated
    • White House
    • Market Stories
    • Election Results
  • Science
  • Sports
  • Technology

    Unlocking AI Success: How People, Process, and Technology Form the Ultimate Triangle

    Billion-dollar coffins? New technology could make oceans transparent and Aukus submarines vulnerable – The Guardian

    Billion-Dollar Coffins? How New Technology Could Make Oceans Transparent and Expose Submarines

    What if artificial intelligence is just a “normal” technology? – The Economist

    What if artificial intelligence is just a “normal” technology? – The Economist

    Lincoln Trail College Receives $100,000 Grant from Marathon Petroleum Corporation for Technology Center – wwbl.com

    Lincoln Trail College Lands $100,000 Grant from Marathon Petroleum to Elevate Technology Center

    Aston Martin to integrate Pirelli’s cyber tyre technology in future models – Just Auto

    Aston Martin to Revolutionize Future Models with Pirelli’s Cutting-Edge Cyber Tyre Technology

    Figure Technology’s stock sizzles after IPO, as investors stay hungry for crypto deals – MarketWatch

    Figure Technology’s Stock Skyrockets After IPO Amid Surging Crypto Investor Excitement

    Trending Tags

    • Nintendo Switch
    • CES 2017
    • Playstation 4 Pro
    • Mark Zuckerberg
No Result
View All Result
  • Home
  • Business
  • Entertainment
    Fall’s Hottest Ski Films Are Going on Tour—Here’s the List – yahoo.com

    Experience the Thrill: Fall’s Hottest Ski Films Hit the Road-Don’t Miss the Tour!

    Ryan Reynolds reveals he called a journalist who said mean things about John Candy – yahoo.com

    Ryan Reynolds Reveals the Moment He Stood Up to a Journalist Who Insulted John Candy

    Entertainment Community Fund Launches Program Supporting Entrepreneurs – Playbill

    Entertainment Community Fund Unveils Exciting New Program to Empower Entrepreneurs

    Behind the turntables: DJ Johnny Kage’s story of perseverance – yahoo.com

    Behind the Turntables: DJ Johnny Kage’s Inspiring Journey of Perseverance

    The other WWE star James Gunn wanted for Peacemaker instead of John Cena – yahoo.com

    The WWE Star James Gunn Originally Wanted for Peacemaker Instead of John Cena

    Quinta Brunson, John Stamos Join Entertainment and Technology Summit – Variety

    Quinta Brunson and John Stamos to Headline Thrilling Entertainment and Technology Summit

  • General
  • Health
  • News

    Cracking the Code: Why China’s Economic Challenges Aren’t Shaking Markets, Unlike America’s” – Bloomberg

    Trump’s Narrow Window to Spread the Truth About Harris

    Trump’s Narrow Window to Spread the Truth About Harris

    Israel-Gaza war live updates: Hamas leader Ismail Haniyeh assassinated in Iran, group says

    Israel-Gaza war live updates: Hamas leader Ismail Haniyeh assassinated in Iran, group says

    PAP Boss to Niger Delta Youths, Stay Away from the Protest

    PAP Boss to Niger Delta Youths, Stay Away from the Protest

    Court Restricts Protests In Lagos To Freedom, Peace Park

    Court Restricts Protests In Lagos To Freedom, Peace Park

    Fans React to Jazz Jennings’ Inspiring Weight Loss Journey

    Fans React to Jazz Jennings’ Inspiring Weight Loss Journey

    Trending Tags

    • Trump Inauguration
    • United Stated
    • White House
    • Market Stories
    • Election Results
  • Science
  • Sports
  • Technology

    Unlocking AI Success: How People, Process, and Technology Form the Ultimate Triangle

    Billion-dollar coffins? New technology could make oceans transparent and Aukus submarines vulnerable – The Guardian

    Billion-Dollar Coffins? How New Technology Could Make Oceans Transparent and Expose Submarines

    What if artificial intelligence is just a “normal” technology? – The Economist

    What if artificial intelligence is just a “normal” technology? – The Economist

    Lincoln Trail College Receives $100,000 Grant from Marathon Petroleum Corporation for Technology Center – wwbl.com

    Lincoln Trail College Lands $100,000 Grant from Marathon Petroleum to Elevate Technology Center

    Aston Martin to integrate Pirelli’s cyber tyre technology in future models – Just Auto

    Aston Martin to Revolutionize Future Models with Pirelli’s Cutting-Edge Cyber Tyre Technology

    Figure Technology’s stock sizzles after IPO, as investors stay hungry for crypto deals – MarketWatch

    Figure Technology’s Stock Skyrockets After IPO Amid Surging Crypto Investor Excitement

    Trending Tags

    • Nintendo Switch
    • CES 2017
    • Playstation 4 Pro
    • Mark Zuckerberg
No Result
View All Result
Earth-News
No Result
View All Result
Home Technology

Maxtext: A simple, performant and scalable Jax LLM

April 24, 2024
in Technology
Maxtext: A simple, performant and scalable Jax LLM
Share on FacebookShare on Twitter

Unit Tests

MaxText is a high performance, highly scalable, open-source LLM written in pure Python/Jax and targeting Google Cloud TPUs and GPUs for training and inference. MaxText achieves high MFUs and scales from single host to very large clusters while staying simple and “optimization-free” thanks to the power of Jax and the XLA compiler.

MaxText aims to be a launching off point for ambitious LLM projects both in research and production. We encourage users to start by experimenting with MaxText out of the box and then fork and modify MaxText to meet their needs.

We have used MaxText to demonstrate high-performance, well-converging training in int8 and scale training to ~51K chips.

Key supported features:

TPUs and GPUs (in preview)
Training and Inference (in preview)
Models: Llama2, Mistral and Gemma

Getting Started
Runtime Performance Results
Comparison To Alternatives
Development
Features and Diagnostics

For your first time running MaxText, we provide specific instructions.

MaxText supports training and inference of various open models. Follow user guides in the getting started folder to know more.

Some extra helpful guides:

Gemma: a family of open-weights Large Language Model (LLM) by Google DeepMind, based on Gemini research and technology. You can run decode and finetuning using these instructions.
Llama2: a family of open-weights Large Language Model (LLM) by Meta. You can run decode and finetuning using these instructions.

In addition to the getting started guides, there are always other MaxText capabilities that are being constantly being added! The full suite of end-to-end tests is in end_to_end. We run them with a nightly cadence. They can be a good source for understanding MaxText Alternatively you can see the continuous unit tests which are run almost continuously.

More details on reproducing these results can be found in MaxText/configs/README.md.

No. of params
Accelerator Type
TFLOP/chip/sec
Model flops utilization (MFU)

32B
v5p-128
3.28e+02
71.47%

64B
v5p-128
3.23e+02
70.31%

128B
v5p-256
3.15e+02
68.68%

128B
v5p-512
3.15e+02
68.53%

256B
v5p-1024
3.16e+02
68.82%

512B
v5p-1024
2.94e+02
63.99%

1024B
v5p-2048
2.49e+02
64.05%

1024B
v5p-4096
2.97e+02
64.80%

1160B
v5p-7680
2.95e+02
64.27%

1160B
v5p-12288
3.04e+02
66.23%

For 16B, 32B, 64B, and 128B models. See full run configs in MaxText/configs/v5e/ as 16b.sh, 32b.sh, 64b.sh, 128b.sh.

Hardware
16B TFLOP/sec/chip
16B MFU
32B TFLOP/sec/chip
32B MFU
64B TFLOP/sec/chip
64B MFU
128B TFLOP/sec/chip
128B MFU

1x v5e-256
120
61.10%
132
66.86%
118
59.90%
110
56.06%

2x v5e-256
117
59.37%
128
64.81%
112
56.66%
110
55.82%

4x v5e-256
117
59.14%
126
64.10%
110
55.85%
108
54.93%

8x v5e-256
115
58.27%
125
63.67%
108
54.96%
104
52.93%

16x v5e-256
111
56.56%
123
62.26%
105
53.29%
100
50.86%

32x v5e-256
108
54.65%
119
60.40%
99
50.18%
91
46.25%

MaxText is heavily inspired by MinGPT/NanoGPT, elegant standalone GPT implementations written in PyTorch and targeting Nvidia GPUs. MaxText is more complex, supporting more industry standard models and scaling to tens of thousands of chips. Ultimately MaxText has an MFU more than three times the 17% reported most recently with that codebase, is massively scalable and implements a key-value cache for efficient auto-regressive decoding.

MaxText is more similar to Nvidia/Megatron-LM, a very well tuned LLM implementation targeting Nvidia GPUs. The two implementations achieve comparable MFUs. The difference in the codebases highlights the different programming strategies. MaxText is pure Python, relying heavily on the XLA compiler to achieve high performance. By contrast, Megatron-LM is a mix of Python and CUDA, relying on well-optimized CUDA kernels to achieve high performance.

MaxText is also comparable to Pax. Like Pax, MaxText provides high-performance and scalable implementations of LLMs in Jax. Pax focuses on enabling powerful configuration parameters, enabling developers to change the model by editing config parameters. By contrast, MaxText is a simple, concrete implementation of various LLMs that encourages users to extend by forking and directly editing the source code.

When running a Single Program, Multiple Data (SPMD) job on accelerators, the overall process can hang if there is any error or any VM hangs/crashes for some reason. In this scenario, capturing stack traces will help to identify and troubleshoot the issues for the jobs running on TPU VMs.

The following configurations will help to debug a fault or when a program is stuck or hung somewhere by collecting stack traces. Change the parameter values accordingly in MaxText/configs/base.yml:

Set collect_stack_trace: True to enable collection of stack traces on faults or when the program is hung. This setting will periodically dump the traces for the program to help in debugging. To disable this, set collect_stack_trace: False.
Set stack_trace_to_cloud: False to display stack traces on console. stack_trace_to_cloud: True will create a temporary file in /tmp/debugging in the TPUs to store the stack traces. There is an agent running on TPU VMs that will periodically upload the traces from the temporary directory to cloud logging in the gcp project. You can view the traces in Logs Explorer on Cloud Logging using the following query:

logName=”projects//logs/tpu.googleapis.com%2Fruntime_monitor”
jsonPayload.verb=”stacktraceanalyzer”

stack_trace_interval_seconds signifies the duration in seconds between each stack trace collection event. Setting stack_trace_interval_seconds: 600 will collect the stack traces every 600 seconds (10 minutes).

Here is the related PyPI package: https://pypi.org/project/cloud-tpu-diagnostics.

Ahead of Time Compilation (AOT, tpu-only)

To compile your training run ahead of time, we provide a tool train_compile.py. This tool allows you to compile the main train_step in train.py for target hardware (e.g. a large number of v5e devices) without using the target hardware, and instead you may use only a CPU or a single VM from a different family. This compilation helps with two main goals:

It will flag any out of memory (OOM) information, such as when the per_device_batch_size is set too high, with an identical OOM stack trace as if it was compiled on the target hardware.

The ahead of time compilation can be saved and then loaded for fast startup and restart times on the target hardware.

The tool train_compile.py is tightly linked to train.py and uses the same configuration file configs/base.yml. Although you don’t need to run on a TPU, you do need to install jax[tpu] in addition to other dependencies, so we recommend running setup.sh to install these if you have not already done so.

Example AOT 1: Compile ahead of time basics

After installing the dependencies listed above, you are ready to compile ahead of time:

# Run the below on a single machine, e.g. a CPU
python3 MaxText/train_compile.py MaxText/configs/base.yml compile_topology=v5e-256 compile_topology_num_slices=2
global_parameter_scale=16 per_device_batch_size=4

This will compile a 16B parameter MaxText model on 2 v5e pods.

Example AOT 2: Save compiled function, then load and run it

Here is an example that saves then loads the compiled train_step, starting with the save:

Step 1: Run AOT and save compiled function

# Run the below on a single machine, e.g. a CPU
export LIBTPU_INIT_ARGS=”–xla_enable_async_all_gather=true”
python3 MaxText/train_compile.py MaxText/configs/base.yml compile_topology=v5e-256
compile_topology_num_slices=2
compiled_trainstep_file=my_compiled_train.pickle global_parameter_scale=16
per_device_batch_size=4 steps=10000 learning_rate=1e-3

Step 2: Run train.py and load the compiled function

To load the compiled train_step, you just need to pass compiled_trainstep_file=my_compiled_train.pickle into train.py:

# Run the below on each host of the target hardware, e.g. each host on 2 slices of v5e-256
export LIBTPU_INIT_ARGS=”–xla_enable_async_all_gather=true”
python3 MaxText/train.py MaxText/configs/base.yml run_name=example_load_compile
compiled_trainstep_file=my_compiled_train.pickle
global_parameter_scale=16 per_device_batch_size=4 steps=10000 learning_rate=1e-3
base_output_directory=gs://my-output-bucket dataset_path=gs://my-dataset-bucket

In the save step of example 2 above we included exporting the compiler flag LIBTPU_INIT_ARGS and learning_rate because those affect the compiled object my_compiled_train.pickle. The sizes of the model (e.g. global_parameter_scale, max_sequence_length and per_device_batch) are fixed when you initially compile via compile_train.py, you will see a size error if you try to run the saved compiled object with different sizes than you compiled with. However a subtle note is that the learning rate schedule is also fixed when you run compile_train – which is determined by both steps and learning_rate. The optimizer parameters such as adam_b1 are passed only as shaped objects to the compiler – thus their real values are determined when you run train.py, not during the compilation. If you do pass in different shapes (e.g. per_device_batch), you will get a clear error message reporting that the compiled signature has different expected shapes than what was input. If you attempt to run on different hardware than the compilation targets requested via compile_topology, you will get an error saying there is a failure to map the devices from the compiled to your real devices. Using different XLA flags or a LIBTPU than what was compiled will probably run silently with the environment you compiled in without error. However there is no guaranteed behavior in this case; you should run in the same environment you compiled in.

Automatically Upload Logs to Vertex Tensorboard

MaxText supports automatic upload of logs collected in a directory to a Tensorboard instance in Vertex AI. Follow user guide to know more.

>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : Hacker News – https://github.com/google/maxtext

Tags: MaxtextSimpletechnology
Previous Post

BeeBase, a programmable relational database with graphical user interface

Next Post

Golang PGO builds using GitHub Actions

Washington takes on flawed federal climate report – NewsRadio 560 KPQ

Washington Challenges Key Flaws in Federal Climate Report

September 14, 2025
James Webb telescope’s ‘starlit mountaintop’ could be the observatory’s best image yet — Space photo of the week – Live Science

James Webb Telescope Unveils Stunning ‘Starlit Mountaintop’ – Perhaps Its Most Breathtaking Image Yet

September 14, 2025

Colby College Announces Exciting New $150 Million Science Facility Funded by Anonymous Donor

September 14, 2025
Maine Coon Cat’s Sweet Love Story With Tiny Sphynx Kitty Is Touching Hearts – yahoo.com

Adorable Maine Coon and Tiny Sphynx Share a Heartwarming Love Story

September 14, 2025

Unlocking AI Success: How People, Process, and Technology Form the Ultimate Triangle

September 14, 2025
Transgender student who challenged Idaho sports ban drops lawsuit – KTVB

Transgender student who challenged Idaho sports ban drops lawsuit – KTVB

September 14, 2025
Ricky Hatton, former world boxing champion, found dead in his home at age 46 – cbsnews.com

Heartbreaking News: Former World Boxing Champion Ricky Hatton Dies at 46

September 14, 2025
US, Chinese officials hold talks in Spain on trade irritants, TikTok deadline – Reuters

US and Chinese Officials Convene in Spain to Resolve Trade Disputes and Address TikTok Deadline

September 14, 2025
Fall’s Hottest Ski Films Are Going on Tour—Here’s the List – yahoo.com

Experience the Thrill: Fall’s Hottest Ski Films Hit the Road-Don’t Miss the Tour!

September 14, 2025
What’s Happening in Health: Kids & Mental Health Town Hall – WKRC

Inside the Kids & Mental Health Town Hall: Key Insights You Can’t Miss

September 14, 2025

Categories

Archives

September 2025
MTWTFSS
1234567
891011121314
15161718192021
22232425262728
2930 
« Aug    
Earth-News.info

The Earth News is an independent English-language daily published Website from all around the World News

Browse by Category

  • Business (20,132)
  • Ecology (820)
  • Economy (839)
  • Entertainment (21,717)
  • General (17,037)
  • Health (9,883)
  • Lifestyle (855)
  • News (22,149)
  • People (843)
  • Politics (848)
  • Science (16,050)
  • Sports (21,340)
  • Technology (15,822)
  • World (822)

Recent News

Washington takes on flawed federal climate report – NewsRadio 560 KPQ

Washington Challenges Key Flaws in Federal Climate Report

September 14, 2025
James Webb telescope’s ‘starlit mountaintop’ could be the observatory’s best image yet — Space photo of the week – Live Science

James Webb Telescope Unveils Stunning ‘Starlit Mountaintop’ – Perhaps Its Most Breathtaking Image Yet

September 14, 2025
  • About
  • Advertise
  • Privacy & Policy
  • Contact

© 2023 earth-news.info

No Result
View All Result

© 2023 earth-news.info

No Result
View All Result

© 2023 earth-news.info

Go to mobile version