issues
search
coreweave
/
tensorizer
Module, Model, and Tensor Serialization/Deserialization
MIT License
180
stars
25
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
ignore
#128
bchess
closed
5 months ago
0
[3.0-dev] Major refactor of serialization in prep for 3.0
#127
bchess
closed
3 months ago
6
feat(docs): Add MkDocs template site
#126
dfinster
closed
2 months ago
2
[Snyk] Security upgrade torch from 1.13.1 to 2.2.0
#125
secgitopscore
opened
5 months ago
0
perf(serialization): Calculate dynamic `num_readers` counts efficiently
#124
Eta0
closed
5 months ago
2
fix(stream_io): Use exclusive end for `CURLStreamFile`s + misc bugs
#123
Eta0
closed
5 months ago
0
feat(serialization): Use a dynamic `num_readers` by default
#122
Eta0
closed
5 months ago
0
feat(stream_io): Remove range request restrictions
#121
Eta0
closed
5 months ago
0
[rough draft] Cython-based file copying; aligned tensors
#120
bchess
opened
6 months ago
0
Jax support
#119
dmarx
opened
6 months ago
0
feat(azure): Preliminary support for Azure
#118
wbrown
opened
6 months ago
0
chore: Update code version to 2.9.0a0, document changes
#117
Eta0
closed
6 months ago
0
Release 2.9.0?
#116
mikeoconnor0308
closed
5 months ago
2
feat(serialization): Add serialization of structured objects
#115
Eta0
closed
6 months ago
1
chore: Update pre-commit hooks
#114
Eta0
closed
6 months ago
0
feat(serialization): Add serializer parameter to limit CPU concurrency
#113
Eta0
closed
6 months ago
0
[Snyk] Fix for 2 vulnerabilities
#112
secgitopscore
opened
6 months ago
0
[Snyk] Fix for 2 vulnerabilities
#111
secgitopscore
opened
6 months ago
0
perf: Add a syscall-based GIL-free address prefaulting function
#110
Eta0
closed
7 months ago
0
fix(serialization): Drop `atomics` dependency & make perf stats optional
#109
Eta0
closed
7 months ago
1
[Snyk] Fix for 2 vulnerabilities
#108
secgitopscore
closed
7 months ago
0
[Snyk] Fix for 2 vulnerabilities
#107
secgitopscore
closed
7 months ago
0
fix(stream_io): Fix `accel-object` parsing with bucket name subdomains
#106
Eta0
closed
7 months ago
0
fix(examples) Fix deserialize timer
#105
bchess
closed
7 months ago
0
perf(serialization): Preload CUDA
#104
Eta0
closed
7 months ago
0
perf(serialization): Prefer `CURLStreamFile._fork` over `open_stream`
#103
Eta0
closed
7 months ago
0
perf(stream_io): Don't use a temporary buffer for truncated reads
#102
Eta0
closed
7 months ago
0
perf(serialization) Make tensors contiguous via thread pool
#101
bchess
closed
6 months ago
2
fix(serialization): Limit `read_numpy_arrays` to CPU deserialization
#100
Eta0
closed
7 months ago
0
fix: Avoid regressions, add tests, better reference handling, misc. logic errors
#99
Eta0
closed
7 months ago
0
perf(serialization) Reuse a pinned buffer when copying from gpu
#98
bchess
closed
7 months ago
2
fix(stream_io): Use region & signature version on S3 writes; use kwargs
#97
Eta0
closed
7 months ago
1
fix(examples): Generate non-persistent buffers on the intended device
#96
Eta0
closed
7 months ago
0
docs: Update `plaid_mode` stats in the README
#95
Eta0
closed
7 months ago
0
fix(redis): relax Redis requirements to `>=4.5.5` from `==5.0.0`.
#94
wbrown
closed
7 months ago
0
support s3 region_name and signature_version in stream_io
#93
lizzzcai
closed
7 months ago
0
support setting S3 region and signature version in stream_io
#92
lizzzcai
closed
7 months ago
0
fix(examples): Move non-persistent buffers to a valid device
#91
Eta0
closed
7 months ago
0
Fail to deserialize model: Expected all tensors to be on the same device
#90
lizzzcai
closed
7 months ago
3
[wip] multireader with mmap allocator
#89
bchess
closed
6 months ago
0
perf(serialization): Use optimal linear partitioning to divide segments
#88
Eta0
closed
7 months ago
0
Major refactor of deserialization to support concurrent readers for performance
#87
bchess
closed
7 months ago
3
tests: Add test for behaviour of out-of-order `pwrite` calls
#86
Eta0
closed
7 months ago
0
perf(serialization): Use Linux `fallocate` instead of `posix_fallocate`
#85
Eta0
closed
7 months ago
0
fix(examples): Ensure using non-cached endpoint when validating
#84
sangstar
closed
7 months ago
0
[WIP] Bchess/multireader merge
#83
bchess
closed
7 months ago
0
fix(serialization): Support serializing tensors over `write` size limits
#82
Eta0
closed
8 months ago
0
Tensorizer Support for Large Models( 70B+) that dont fit into a single GPU
#81
SujoyDutta
opened
8 months ago
11
[WIP] multireader
#80
bchess
closed
7 months ago
3
feat(serialization): Default to `plaid_mode=True` for CUDA tensors
#79
Eta0
closed
8 months ago
0
Previous
Next