* This change avoids crashes when running T5 models with F16 tensors on CPU.
* This enables running ProstT5's (https://huggingface.co/Rostlab/ProstT5) encoder-only mode in Candle. This ProstT5 mode stores it's embed_tokens weights within the encoder, as its decoding stage was replaced with a CNN. You could write more, like: This alone is not sufficient to run ProstT5 within Candle examples. We will develop a ProstT5 runner outside candle for now, but would be willing to upstream it to candle-examples at a later point.
* Revert "This enables running ProstT5's (https://huggingface.co/Rostlab/ProstT5) encoder-only mode in Candle. This ProstT5 mode stores it's embed_tokens weights within the encoder, as its decoding stage was replaced with a CNN. You could write more, like: This alone is not sufficient to run ProstT5 within Candle examples. We will develop a ProstT5 runner outside candle for now, but would be willing to upstream it to candle-examples at a later point."
This reverts commit d886d3ce5e.
* This change avoids crashes when running T5 models with F16 tensors on CPU.
* This enables running ProstT5's (https://huggingface.co/Rostlab/ProstT5) encoder-only mode in Candle. This ProstT5 mode stores it's embed_tokens weights within the encoder, as its decoding stage was replaced with a CNN. This alone is not sufficient to run ProstT5 within Candle examples. We will develop a ProstT5 runner outside candle for now, but would be willing to upstream it to candle-examples at a later point.
* Start adding the recurrent-gemma model.
* More griffin.
* Add the example + get the weights to load from the HF version.
* More inference code.
* Rope + kv-cache on the attention side.
* Add to the inference code.
* Add more to the recurrent gemma inference.
* Get some first inference to run.
* Add the softcap on logits.
* Fixes.
* Use partial rotary embeddings.
* Get inference to work.
* Add a comment.
* And add a readme.
* Move the metal kernels utils in a separate module.
* Use the BufferOffset for unary ops.
* Fix clippy lints.
* Use the new BufferOffset.
* Adapt the binary ops.
* Affine.
* More ops (powf, elu, cast).
* moondream implementation
* add moondream example
* change config default activation
* Add assets and integrate phi mixformer with example
* Make use of kv cache and fix seq_len bug; Clean up example code
* Add README link to example
* Remove pos_embed scaling; Remove assets; Add to README; Expand VisionConfig
* Delete image
* Use apply instead of forward
* Use latest release special token; Fix token/s accuracy; Use GeluPytorchTanh in VisionConfig v2
* Add flag to use f16
* Avoid breaking the quantized version on cuda.
---------
Co-authored-by: laurent <laurent.mazare@gmail.com>
* add the sign unary operator
* remove uneeded import
* remove uneeded import
* undo formatting
* undo formatting
* remove unnecessary redefintion
* allow gradient to flow through for sign and round
* fix cpu ops to ensure that negzero and positive zero are handled properly
* clippy fixes
* Properly avoid gradient tracking.
* Use a branchless version.
---------
Co-authored-by: laurent <laurent.mazare@gmail.com>
* moondream implementation
* add moondream example
* change config default activation
* Add assets and integrate phi mixformer with example
* Make use of kv cache and fix seq_len bug; Clean up example code
* Add README link to example
* Remove pos_embed scaling; Remove assets; Add to README; Expand VisionConfig
* Delete image
* Use apply instead of forward
* Use latest release special token; Fix token/s accuracy; Use GeluPytorchTanh in VisionConfig v2
* moondream implementation
* add moondream example
* change config default activation
* Add assets and integrate phi mixformer with example
* Make use of kv cache and fix seq_len bug; Clean up example code
* Add README link to example
* Remove pos_embed scaling; Remove assets; Add to README; Expand VisionConfig
* Delete image
* Use apply instead of forward
* Pass bos token at the beginning of tensor.
* Quantize moondream.
* Forward with image bos token.
* Clippy.
* Use q4_0 quantization.
* Add pointers for sequence and tokens; Remove seq_len conditional
* Add more cuda kernels for quantized matmul.
* Add the vec-dot bits.
* Expose the quantized matmul-vec kernels.
* Also include the quantize-q8-1 kernel.
* Glue code for the q8-1 quantization.
* mm-vec product via q8-1 quantization.
* Add a test.
* Add a mm test.
* Get the test to return some sensible results.
* Also test dmmv.
* Fix the launch params.
* Allow for tweaking the force_dmmv parameter while it's experimental.
