* Pip: Move from tensorflow-gpu to tensorflow package
Since tensorflow 2.1 the regular tensorflow package contains GPU support and replaces the tensorflow-gpu package, which will be depreceated at some point.
* Readme: Update Tensorflow version
* Update Pipfile.lock
Co-authored-by: Willem Melching <willem.melching@gmail.com>
old-commit-hash: 3f8afe593f14a4a6f94ddb9c2f1aed1419d2f69a
* Create workflow to update the Pipfile.lock weekly
* Update-pipfile: Also install wheel
* Install wheel in different step
* Pipfile: Only lock major versions
Assumes Semantic Versioning, and only sets the major versions of all packages
* Run update-pipfile on changes in pipfile and use tensorflow
Move from tensorflow-gpu to tensorflow, an excelent test for this PR
* [create-pull-request] automated change
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
old-commit-hash: b1795235f9767ac38f1bc8c7cd2dee9d4f646cf6
* only send requests for the right brand
* this works on sonata
Co-authored-by: openpilot laptop <laptop@comma.ai>
old-commit-hash: 1162041ea63a7213b9e2b9c11edd521c89f86c99
* handle posenet and senor alerts in locationd
* defaults now set in capnp file
* Cleanup c++ version of params learner
* update ref commit
old-commit-hash: d099e09fb77544eb00d13ba9aea34a0a3374a52e
* hyundai fw query
* Change query
* this works
* That is not engine
* hyundai fw query
* Change query
* this works
* That is not engine
* Skip FW query in test_car_models
* Those routes don't fingerprint properly after speedup
* only send toyota queries to subaddresses
Co-authored-by: openpilot laptop <laptop@comma.ai>
old-commit-hash: bb8bbdd9ee4d610e2ae6048021407863b3f74683
* don't compile in docker build stage
* disable that for now
* do we need that?
* upload artifact already compresses
* that wasn't faster
This reverts commit 2221e13c71b5cc693dd3bebfa4061d4ab6821437.
* try that
* don't think that's used
* this too
* check this
* less layers
* that's not in the dockerfile anymore
* system
* even less layers and clear pipenv cache
* no inline comment i guess
* uninstall pipenv after we're done with it
* separate build job adds too much overhead
* quotes
* clean that up
* increase timeout
* cleanup
old-commit-hash: 45ee131a4427db6d97531849bf4a1c3206ecbd19
* add thneed self test
* don't do the memset in thneed, shouldn't matter though
Co-authored-by: Comma Device <device@comma.ai>
old-commit-hash: 6c0ad1e67567e149cdf9e00ab7ae43f4cc5bc579
* bigmodel
* more debug print
* debugging bigmodel
* remove the tanh, debugging
* print images/buffers
* disassemble the command queues
* decompiler
* dump the shaders
* full disasm
* support patching kernel and fixing convolution_horizontal_reduced_reads_1x1
* microbenchmark
* 42 GFLOPS, 1 GB/s
* gemm benchmark
* 75 GFLOPS vs 42 GFLOPS
* 115 GFLOPS
* oops, never mind
* gemm image is slow
* this is pretty hopeless
* gemm image gets 62 GFLOPS
* this is addictive and still a waste of time
* cleanup cleanup
* that hook was dumb
* tabbing
* more tabbing
Co-authored-by: Comma Device <device@comma.ai>
old-commit-hash: 78a352a8ca8a948e86e7c752732e470f89d92280
* way cleaner take 2
* cleanup
* be more relaxed
* just let it be
* don't drive backwards or upside down
* do this more
* vNED sometyimes invalid
* use reasonble stds
* stability in nonlinear zone
* previous metrics were without publishing
old-commit-hash: 81686547cc10dc129fb921f938dc496455613e33
Remove panda / giraffe mention in footnotes where possible
Car harnesses now available, removed subscript for known HKG models; left subscript for not known models
old-commit-hash: 5677f45a655b1d065f77f55e37b300e10dadcfd0
Willem Melching