tenstorrent tt-forge-fe issues

tenstorrent / tt-forge-fe

The TT-Forge FE is a graph compiler designed to optimize and transform computational graphs for deep learning models, enhancing their performance and efficiency.

https://docs.tenstorrent.com/tt-forge-fe/

Apache License 2.0

19 stars 3 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

FORGE_DISABLE_ERASE_INVERSE_OPS_PASS flag cleanup and refactor abs,exp,maximum,less,greater,not_equal ops test cases

#605 kamalrajkannan78 opened 1 day ago
0
[Model Analysis] Create sample that reproduces MLIR seg fault

#604 nvukobratTT opened 2 days ago
0
Script to run generated standalone Forge tests + report summary

#603 nvukobratTT opened 2 days ago
0
[Model Analysis] Create script to run standalone tests and generate output report

#602 nvukobratTT opened 2 days ago
0
Option to setup compiler configurations using pytest arguments

#601 nvukobratTT opened 2 days ago
0
[Model Analysis] Define option to use pytest argument to setup compiler configurations

#600 nvukobratTT opened 2 days ago
0
Generate op tests based on run model

#599 nvukobratTT opened 2 days ago
0
[Model Analysis] Filter out op tests to generate only unique ones

#598 nvukobratTT opened 2 days ago
0
TMS exist on graph edges from initial graph

#597 dgolubovicTT opened 2 days ago
1
New Convolution op in MLIR. Need to change lowering of Conv2d

#596 LPanosTT opened 2 days ago
0
Expand TVM Codegen functionality to enable generation of pytest functions withing the module.

#595 nvukobratTT opened 2 days ago
0
[Model Analysis] Expand TVM Codegen to support pytest function generation

#594 nvukobratTT opened 2 days ago
0
Store node inputs during TVM codegen op parsing

#593 nvukobratTT opened 2 days ago
0
[Model Analysis] Store operation input shapes during TVM codegen

#592 nvukobratTT opened 2 days ago
0
Remove the requirement to push the whole Framework Module as input to the Forge Module during compile time

#591 nvukobratTT opened 2 days ago
0
[Model Analysis] Remove the need of using Framework module within generated Forge module

#590 nvukobratTT opened 2 days ago
0
[Model Analysis] Utilize TVM Codegen to generate op tests

#589 nvukobratTT opened 3 days ago
0
[Model Analysis] Based on generated model op tests, create summary for all ops

#588 nvukobratTT opened 3 days ago
0
[Model Analysis] Based on generated model op tests, gather results about passing/failing states

#587 nvukobratTT opened 3 days ago
0
[Model Analysis] Generate op tests based on analysis results of the model

#586 nvukobratTT opened 3 days ago
0
[Model Analysis] Analyze model when entering TVM

#585 nvukobratTT opened 3 days ago
0
[Model Analysis] Export .xls files next to CSV files

#584 nvukobratTT opened 3 days ago
0
[Op] Test conv2d transpose (deconv) op

#583 nvukobratTT opened 3 days ago
0
[Op] Test where op

#582 nvukobratTT opened 3 days ago
0
[Op] Test clip op

#581 nvukobratTT opened 3 days ago
0
[Op] Test gelu op

#580 nvukobratTT opened 3 days ago
0
[Op] Test layernorm op

#579 nvukobratTT opened 3 days ago
0
[Op] Test batchnorm

#578 nvukobratTT opened 3 days ago
0
[Op] Test leaky_relu op

#577 nvukobratTT opened 3 days ago
0
[Op] Test tanh op

#576 nvukobratTT opened 3 days ago
0
[Op] Test cosine op (ttir.cos)

#575 nvukobratTT opened 3 days ago
0
[Op] Test sin op (ttir.sin)

#574 nvukobratTT opened 3 days ago
0
[Op Testing] ne, gt, lt, maximum, minimum

#573 vbrkicTT opened 3 days ago
0
Bfloat16 weights are not recognized by tvm

#572 ndrakulicTT opened 3 days ago
1
[Docs] TT-Forge-FE compiler architecture overview

#571 nvukobratTT opened 3 days ago
0
Document TT-Forge-FE architecture

#570 nvukobratTT opened 3 days ago
0
Upload perf report

#569 vmilosevic opened 4 days ago
0
Fix for passing unnecessary inputs to the backward pass of the

#568 ndrakulicTT closed 3 days ago
0
Fix decompose of AvgPool2D to account for named attributes and skip runtime reinterpreting shapes

#567 dgolubovicTT closed 3 days ago
0
Fix hoisting transpose before broadcasts

#566 dgolubovicTT closed 2 days ago
2
Broadcast TMs are improperly commuted through transpose op

#565 dgolubovicTT closed 2 days ago
0
Increase package level for development docker images

#564 nvukobratTT opened 4 days ago
1
TVM casts all bfloat16 data formats to float32

#563 dgolubovicTT opened 4 days ago
0
Add option for dumping the unique op config as xlsx file and script to create sheet for model unique op configs

#562 chandrasekaranpradeep opened 4 days ago
0
Update submodules for docker build

#561 vmilosevic closed 3 days ago
0
Attribute 'keep_dim' error in 'ttir.mean' op in adaptive_avg_pool_2d

#560 kamalrajkannan78 closed 3 days ago
1
Attribute 'shape' error in 'ttir.reshape' op in adaptive_avg_pool_2d

#559 kamalrajkannan78 closed 3 days ago
1
Documentation for single operator E2E tests

#558 vbrkicTT opened 5 days ago
0
Update tvm reference

#557 dgolubovicTT closed 4 days ago
1
Tests for concatenate op

#556 kmilanovicTT opened 5 days ago
0