index
:
specification.git
main
v0.23
v0.30
v0.40
v0.50
v0.60
v0.70
v0.80
v0.90
v1.00
[no description]
about
summary
refs
log
tree
commit
diff
log msg
author
committer
range
Age
Commit message (
Expand
)
Author
2024-03-19
Remove draft tag from 0.90 branch
v0.90
Eric Kunze
2024-01-31
Add section of shape operators
v0.90.0
Eric Kunze
2024-01-28
Add missing pseudocode for CONST, CUSTOM, ERF, and IDENTITY operators
Kevin Petit
2024-01-25
Handle CAST from fp-type to narrow integer types
Eric Kunze
2024-01-24
pseudocode: replace all uses of FP64 and double types with fp64_t
Kevin Petit
2024-01-23
Fix syntax higlighting of pseudocode
Kevin Petit
2024-01-23
pseudocode: use correct variable to write out data in REDUCE_MIN
Kevin Petit
2024-01-23
pseudocode: fix name mismatches with operator arguments in the XML
Kevin Petit
2024-01-23
pseudocode: replace C-style casts with static_cast
Kevin Petit
2024-01-23
pseudocode: use true/false consistently for boolean values
Kevin Petit
2024-01-19
Fix typo in argmax pseudocode
Thomas Preud'homme
2024-01-19
pseudocode: only use commas to separate ranges in for_each statements
Kevin Petit
2024-01-16
pseudocode: a collection of trivial fixes
Kevin Petit
2024-01-15
Move operator pseudocode to separate files
Kevin Petit
2024-01-15
Main Conformance: TANH change, SCATTER clarification
Dominic Symes
2024-01-09
Fix typo in i48_t description
Eric Kunze
2024-01-09
Main Conformance: Clarify REDUCE_PRODUCT conformance
Dominic Symes
2024-01-08
Add note to POW accuracy
Eric Kunze
2023-12-15
IDENTITY: Add missing types
Dominic Symes
2023-12-04
Main Conformance: Update SIGMOID, TANH precision
Dominic Symes
2023-11-28
REDUCE_SUM: Specify the accumulator type
Dominic Symes
2023-11-22
Main Conformance: Update RSQRT precision
Dominic Symes
2023-11-03
Fix description of CONST output
Eric Kunze
2023-11-02
Fix CustomOp description text
Jerry Ge
2023-11-01
Update CustomOp's attribute names
Jerry Ge
2023-11-01
Main Conformance: Update EXP, POW precision
Dominic Symes
2023-10-30
Add description of bias broadcast
Eric Kunze
2023-10-09
Main conformance: clarify floating point accuracy
Dominic Symes
2023-09-26
Define missing accuracy bounds
Eric Kunze
2023-09-21
TRANSPOSE_CONV2D: Remove redundant out_shape argument
Dominic Symes
2023-09-13
Update version to 0.90 draft
Eric Kunze
2023-09-12
Main Conformance: Add local_bound to RFFT2D
Dominic Symes
2023-09-11
Add integer divide with floor for coordinate calculation
Eric Kunze
2023-09-07
Main Conformance: Update dot product error bound
v0.90a0
Dominic Symes
2023-08-18
Update Custom operator arguments
Eric Kunze
2023-08-17
Change TOSA specification to signless types
Eric Kunze
2023-08-16
Main conformance: Add local bound flag for convolutions
Dominic Symes
2023-08-16
Add DIM operator and operations on shape_t values
Dominic Symes
2023-08-15
Increase MAX_SCALE for the 8K level
Eric Kunze
2023-08-11
Add StatefulOps to TOSA specification
Jerry Ge
2023-08-07
Allow bias tensors to be broadcast
Dominic Symes
2023-08-03
Use black and flake8 to format python (NFC)
Eric Kunze
2023-07-24
COND_IF/WHILE_LOOP: Add nesting level limit
Dominic Symes
2023-07-11
Change max rank for reduction operators to MAX_RANK
Eric Kunze
2023-07-06
AVG_POOL2D conformance: fix kernel size names
Dominic Symes
2023-06-23
Add Int4 to ConstOp supported data types
Jerry Ge
2023-06-23
FULLY_CONNECTED: Change weights and bias to be inputs
Dominic Symes
2023-06-21
Align RESCALE zp description with pseudocode
Eric Kunze
2023-06-21
Fix ErfOp to match with the rest of activation operators
Jerry Ge
2023-06-21
Add ERF activation function
Eric Kunze
[next]