Lukas Kreussel c72eb3d75b Add reference implementation for q4k and q5k (#586) ...

* add `q2k` vec-dot

* `q3k` vec-dot + quantization bugfix

* `q4k` vec-dot

* `q5k` vec-dot

* Validate against GGML unit test results.

* Remove some more `transmutes`

2023-08-26 12:07:54 +01:00

..

conv_tests.rs

Add to the cuda example a reproduction of the issue. (#579)

2023-08-24 12:07:31 +01:00

custom_op_tests.rs

Relax the requirements on CustomOp. (#486)

2023-08-17 11:12:05 +01:00

display_tests.rs

Rename the candle crate to candle-core (#301)

2023-08-02 08:20:22 +01:00

grad_tests.rs

Fix the minimum/maximum gradient computations. (#534)

2023-08-21 08:28:41 +01:00

indexing_tests.rs

Rename the candle crate to candle-core (#301)

2023-08-02 08:20:22 +01:00

layout_tests.rs

Rename the candle crate to candle-core (#301)

2023-08-02 08:20:22 +01:00

pool_tests.rs

More non square testing (#582)

2023-08-24 13:01:04 +01:00

quantized_tests.rs

Add reference implementation for q4k and q5k (#586)

2023-08-26 12:07:54 +01:00

tensor_tests.rs

Add a yolo-v3 example. (#528)

2023-08-20 18:19:37 +01:00

test_utils.rs

More accelerate optimizations (#427)

2023-08-13 12:53:34 +01:00