Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / geohot/tinygrad issues and pull requests

#780 - Contract float4 ALU operations

Pull Request - State: closed - Opened by geohot over 1 year ago

#779 - "unbiased" std and Bessel's Correction for BatchNorm2d

Pull Request - State: open - Opened by symlons over 1 year ago

#778 - add bounded_randn func.

Pull Request - State: closed - Opened by rabiaedayilmaz over 1 year ago - 1 comment

#777 - Zero dim Tensor support

Pull Request - State: open - Opened by Joqsan over 1 year ago - 2 comments

#776 - Rdna

Pull Request - State: closed - Opened by geohot over 1 year ago

#775 - UNet3D: Compute Mean DICE score on KiTS19

Pull Request - State: open - Opened by jla524 over 1 year ago - 1 comment

#774 - BatchNorm2D Bessel's correction

Pull Request - State: closed - Opened by symlons over 1 year ago - 2 comments

#773 - Inference test: add tests for ResNet50

Pull Request - State: closed - Opened by jla524 over 1 year ago

#772 - conv3d + unet3d

Pull Request - State: closed - Opened by geohot over 1 year ago

#771 - added leaky relu

Pull Request - State: closed - Opened by rabiaedayilmaz over 1 year ago - 1 comment

#770 - ResNet: match implementation with Nvidia and PyTorch

Pull Request - State: closed - Opened by jla524 over 1 year ago - 1 comment

#769 - Statistics test: check if distributions match torch

Pull Request - State: closed - Opened by jla524 over 1 year ago - 2 comments

#768 - Added more tests

Pull Request - State: closed - Opened by AbhiroopAjith over 1 year ago - 1 comment

#767 - add std to tensor.py

Pull Request - State: closed - Opened by rabiaedayilmaz over 1 year ago - 17 comments

#766 - make CIFAR training faster

Pull Request - State: closed - Opened by geohot over 1 year ago

#765 - 0D tensor support: can it happen?

Issue - State: open - Opened by Joqsan over 1 year ago - 2 comments

#764 - No Copy

Pull Request - State: closed - Opened by geohot over 1 year ago - 2 comments

#763 - multidevice works

Pull Request - State: closed - Opened by geohot over 1 year ago

#762 - multi cl_queue

Pull Request - State: closed - Opened by geohot over 1 year ago

#761 - Fixed package description

Pull Request - State: closed - Opened by rylan-justice over 1 year ago

#760 - Updated LICENSE year

Pull Request - State: closed - Opened by rylan-justice almost 2 years ago

#759 - support multiple devices

Pull Request - State: closed - Opened by geohot almost 2 years ago

#758 - Add Tensor.stack() and Tensor.repeat() (...trying to make einops work with tinygrad)

Pull Request - State: closed - Opened by Joqsan almost 2 years ago - 1 comment

#757 - Updated LICENSE and setup.py

Pull Request - State: closed - Opened by rylan-justice almost 2 years ago - 1 comment

#756 - Added kaiming_uniform initialization for Conv2d and Linear layers

Pull Request - State: open - Opened by rabiaedayilmaz almost 2 years ago - 2 comments

#755 - LR Schedulers

Pull Request - State: open - Opened by mattismegevand almost 2 years ago - 2 comments

#754 - Fix missing dependency for examples/llama.py

Pull Request - State: closed - Opened by Perryman almost 2 years ago - 3 comments

#753 - fix batchnorm at training

Pull Request - State: closed - Opened by geohot almost 2 years ago

#751 - Test the optimizer with JIT

Issue - State: open - Opened by geohot almost 2 years ago

#750 - HIP backend

Pull Request - State: open - Opened by nanamiwang almost 2 years ago - 8 comments

#749 - Fixed two errors when running LLaMA example with CUDA=1

Pull Request - State: closed - Opened by Sense545 almost 2 years ago

#748 - Remove Zeroview

Pull Request - State: closed - Opened by geohot almost 2 years ago

#747 - Fused Optimizer Kernels

Issue - State: open - Opened by geohot almost 2 years ago

#746 - fix backward convs

Pull Request - State: closed - Opened by geohot almost 2 years ago

#745 - Kernel Optimization

Issue - State: open - Opened by geohot almost 2 years ago

#744 - Fix train_resnet example

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#743 - Use tokens as the 'tiny' metric instead of lines

Issue - State: open - Opened by langfield almost 2 years ago

#742 - AttributeError: 'Transformer' object has no attribute 'model'

Issue - State: open - Opened by CRCODE22 almost 2 years ago

#741 - Do not specify errors when trying to load devices.

Pull Request - State: closed - Opened by jintuzhang almost 2 years ago

#740 - padding changes for onnx

Pull Request - State: closed - Opened by sohaiberrabii almost 2 years ago - 1 comment

#739 - Add Support for quantized (4-bit) Models

Issue - State: open - Opened by apcameron almost 2 years ago

#738 - make maximum split the grad like torch when equal

Pull Request - State: closed - Opened by worldwalker2000 almost 2 years ago - 6 comments

#737 - rust backend basics

Pull Request - State: open - Opened by benmkw almost 2 years ago - 9 comments

#736 - Better Local Support

Issue - State: open - Opened by geohot almost 2 years ago

#735 - C Backend for training on embedded devices

Issue - State: open - Opened by benmkw almost 2 years ago - 1 comment

#734 - Uopsif

Pull Request - State: open - Opened by geohot almost 2 years ago

#733 - Device: move below LazyBuffer

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#732 - Device: make GPU default and METAL/CUDA if possible

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 1 comment

#731 - Permute examples

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#730 - good changes from the M1 Tensor Core project

Pull Request - State: closed - Opened by geohot almost 2 years ago

#729 - Added activation ops + tests

Pull Request - State: closed - Opened by AndreSlavescu almost 2 years ago - 4 comments

#728 - Use the Tensor Cores on M1+

Pull Request - State: open - Opened by geohot almost 2 years ago

#726 - casting: support all backends and implement half

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#725 - Check for loops on reduce+elementwise fusion

Issue - State: open - Opened by geohot almost 2 years ago

#724 - Add tests for casting

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 5 comments

#723 - fix binop, other tests failure

Pull Request - State: closed - Opened by geohot almost 2 years ago

#721 - remove the stupid register class

Pull Request - State: closed - Opened by geohot almost 2 years ago

#720 - types

Pull Request - State: closed - Opened by geohot almost 2 years ago

#719 - Remove Register class and use the end of the shapetracker.

Issue - State: closed - Opened by geohot almost 2 years ago

#718 - I keep getting this error

Issue - State: closed - Opened by Asory2010 almost 2 years ago - 7 comments

#717 - ./tinygrade/examples/llama.py editing TOKENIZER_FILENAME & Weights issues

Issue - State: closed - Opened by kencom999 almost 2 years ago - 5 comments

#716 - add AdamW optimizer

Pull Request - State: closed - Opened by croumegous almost 2 years ago - 3 comments

#715 - Remove completed TODOs

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 3 comments

#714 - linearizer!

Pull Request - State: closed - Opened by geohot almost 2 years ago - 1 comment

#713 - Fix constant folding

Pull Request - State: closed - Opened by geohot almost 2 years ago

#712 - Use safetensors in llama

Pull Request - State: closed - Opened by python273 almost 2 years ago - 8 comments

#711 - Use pathlib

Pull Request - State: closed - Opened by python273 almost 2 years ago

#710 - Fix llama 13B RAM usage

Pull Request - State: closed - Opened by python273 almost 2 years ago - 2 comments

#709 - pip install "metal" issue ticket

Issue - State: closed - Opened by u1npacker almost 2 years ago - 7 comments

#708 - Devicebufferless

Pull Request - State: closed - Opened by geohot almost 2 years ago

#707 - Remove extra layer of abstraction

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 4 comments

#706 - docs: Update path for test_mnist in README

Pull Request - State: closed - Opened by connor-henderson almost 2 years ago

#705 - A fix for an import error in docs

Pull Request - State: closed - Opened by pasxn almost 2 years ago

#704 - LLaMa MEM_OBJECT_ALLOCATION_FAILURE when start

Issue - State: closed - Opened by CoffeeAddictGuy almost 2 years ago - 1 comment

#703 - Cast Tensor data to float32

Pull Request - State: closed - Opened by ayushmankumar7 almost 2 years ago - 4 comments

#702 - Cast image to float32

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#701 - KeyError: dtype('uint8') On running Examples

Issue - State: closed - Opened by ayushmankumar7 almost 2 years ago - 2 comments

#700 - Fix llama 13B weights loading

Pull Request - State: closed - Opened by python273 almost 2 years ago - 5 comments

#699 - add int64 as supported dtype from numpy

Pull Request - State: closed - Opened by fernavid almost 2 years ago - 2 comments

#698 - Remove extra layer of abstraction

Issue - State: closed - Opened by geohot almost 2 years ago - 1 comment

#697 - Move rawbuffer

Pull Request - State: closed - Opened by geohot almost 2 years ago

#696 - LazyOp and LazyBuffer are confusing

Issue - State: closed - Opened by geohot almost 2 years ago

#695 - Unify codegen/gpu.py and codegen/llvm.py with the Linearizer

Issue - State: closed - Opened by geohot almost 2 years ago - 4 comments

#694 - Proper (lazy) casting

Issue - State: closed - Opened by geohot almost 2 years ago - 1 comment

#693 - uint8 + int8 support

Issue - State: open - Opened by geohot almost 2 years ago

#692 - SpecializedJit

Issue - State: open - Opened by geohot almost 2 years ago - 1 comment

#691 - Remove ZeroView from ShapeTracker

Issue - State: closed - Opened by geohot almost 2 years ago

#689 - Type cast

Pull Request - State: closed - Opened by dc-dc-dc almost 2 years ago - 2 comments

#687 - LLaMA 13B not running on 32GB M1

Issue - State: closed - Opened by AlexAndrei98 almost 2 years ago - 7 comments

#685 - LLaMA: AttributeError: type object 'CPUBuffer' has no attribute 'empty'

Issue - State: closed - Opened by GrahamboJangles almost 2 years ago - 7 comments

#672 - WASM URL fetching doesn't work

Issue - State: open - Opened by wkoszek almost 2 years ago - 1 comment

#660 - Remove examples/deep_deterministic_policy_gradient.py

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 5 comments

#658 - Failing opt test (rerunning binop)

Pull Request - State: closed - Opened by geohot almost 2 years ago

#646 - OpenCL mad

Pull Request - State: closed - Opened by geohot almost 2 years ago

#617 - replace a lot of the confusing conv reshapes, etc... with einops

Issue - State: open - Opened by geohot almost 2 years ago - 6 comments

#551 - optim tests are only testing one step!

Issue - State: closed - Opened by geohot almost 2 years ago - 3 comments
Labels: good first issue

#549 - Show where tinygrad is faster in speed test vs torch

Pull Request - State: closed - Opened by jla524 almost 2 years ago

#540 - Fix examples

Pull Request - State: closed - Opened by jla524 almost 2 years ago - 3 comments

#472 - Support for Coral devices with Edge TPU

Issue - State: closed - Opened by pythops about 2 years ago - 2 comments