a83ca2ece0
Bump the crate version to 0.4.0. ( #1658 )
2024-02-04 19:08:01 +01:00
153c940a9c
Update docs to reflect current usage of example ( #1610 )
...
modified: candle-examples/examples/onnx/README.md
2024-02-04 11:59:47 +01:00
50be8a98ba
Quantized support for stable-lm2. ( #1654 )
...
* Quantized support for stable-lm2.
* Quantized support for v2-zephyr.
2024-02-04 11:57:05 +01:00
58cc896e69
make llama derive clone ( #1648 )
...
Co-authored-by: danielclough <danielclough@users.noreply.github.com >
2024-02-04 11:56:03 +01:00
5cdd84e0f6
onnx: add the Flatten operator. ( #1638 )
...
* onnx: add the Flatten operator.
* onnx flatten: merge axis condition
---------
Co-authored-by: 王泽龙 <wangzelong@shenqishen.com >
2024-02-03 16:28:47 +01:00
a510ddec4e
Mention the new models in the readme. ( #1651 )
2024-02-03 15:19:57 +01:00
d32abbce53
Add StableLM-2, StableLM Code and Zephyr variants ( #1650 )
...
* Add StableLM Code and Zephyr variants
* Add V2 models
* Update README
2024-02-03 14:58:41 +01:00
dfab45e1c8
Supports more audio formats ( #1628 )
...
* Supports more audio formats
* Simplify the handling of the different buffer types.
* Check the sample rate.
---------
Co-authored-by: laurent <laurent.mazare@gmail.com >
2024-02-03 14:26:04 +01:00
96bc704d17
Update mixformer.rs ( #1601 )
...
Update the source of the configuration_mixformer_sequential.py
It has been removed, therefore, it is still available in this -> d38e6f954ec29b96fe2cf033937dad64e279b5d9
2024-02-03 13:42:16 +01:00
a52d407ae6
Add ConvNeXt model. ( #1604 )
2024-02-03 13:34:28 +01:00
9e824ec810
Explicit version for packages that are not in the workspace. ( #1642 )
2024-01-31 18:57:38 +01:00
beadb1b434
Explicit candle version so that cargo publish can be used easily. ( #1641 )
2024-01-31 18:42:22 +01:00
6d83d42efb
Merge pull request #1606 from FL33TW00D/feature/larger-batches
...
fix: larger batches
2024-01-29 15:31:10 +00:00
b6afb46601
chore: final
2024-01-22 15:15:19 +00:00
fd7c856564
Merge pull request #1533 from huggingface/ivarflakstad/metal-prng
2024-01-22 07:30:20 +01:00
73d79e6092
chore: actual fix
2024-01-19 09:35:42 +00:00
b1879f17f6
chore: switch to buffer
2024-01-19 08:57:49 +00:00
4f79f5df8a
fix: larger batches
2024-01-18 14:30:14 +00:00
1cf34368b7
Merge pull request #1602 from mimiquate/fix-metal-kernel-type
...
Metal: Use uint8_t as output type in int64_t binary op kernel
2024-01-18 08:40:34 +01:00
17e6e2d7ee
Fixes metal kernel u8 type
2024-01-17 15:47:08 -03:00
80b1c689f9
Revert public EncoderParam
2024-01-17 18:09:28 +01:00
db923517b3
Merge branch 'main' into ivarflakstad/metal-prng
2024-01-17 18:03:57 +01:00
403680f17d
Quantized GGUF style ( #1523 )
...
* Metal quantized modifications proposal.
- Add a device param, wherever needed.
- Create new QMetal storage thing that implements QuantizedType.
- Update everywhere needed.
Fix Python.
Fixing examples.
Fix: fmt + clippy + stub.
Moving everything around.
Only missing the actual implems.
Fixing everything + adding dequantized kernels.
More work.
Fixing matmul.
Fmt + Clippy
Some clippy fixes.
Working state.
Q2K Metal -> Bugged (also present in GGML).
Q4K CPU -> Bugged (present previously, new test catch it).
Q5K CPU -> Bugged (present previously).
Q8_1 Both -> Never really implemented it seems
Q8K metal -> Never implemented in metal
Fixing Q2K bug (present in ggml).
* Cleanup.
* Fix the rebase.
* Removing the fences speeds everything up and *is* correct this time...
* Cleanup the fence.
* After rebase.
* Bad code removal.
* Rebase after phi2 merge + fix replit default to CPU.
* Making the CI happy.
* More happy tests.
---------
Co-authored-by: Nicolas Patry <nicolas@Nicolass-MacBook-Pro.local >
2024-01-17 10:27:58 +01:00
86a8e58897
Update metal random kernel and set_seed method
...
* set_seed via buffer content pointer copy + did_modify_range
* ensure random.metal kernel does not write outside of buffer range when tid==0
2024-01-17 09:12:44 +01:00
5270224f40
Add MobileOne model. ( #1595 )
...
* Add MobileOne model.
* Clippy fixes
* Remove a comment.
---------
Co-authored-by: laurent <laurent.mazare@gmail.com >
2024-01-16 06:34:16 +01:00
7e3349d7c3
Update parquet requirement from 45.0.0 to 50.0.0 ( #1592 )
...
Updates the requirements on [parquet](https://github.com/apache/arrow-rs ) to permit the latest version.
- [Changelog](https://github.com/apache/arrow-rs/blob/master/CHANGELOG-old.md )
- [Commits](https://github.com/apache/arrow-rs/compare/45.0.0...45.0.0 )
---
updated-dependencies:
- dependency-name: parquet
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com >
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-01-15 22:35:01 +01:00
1257fc6719
Update safetensors requirement from 0.3.1 to 0.4.1 ( #1591 )
...
Updates the requirements on [safetensors](https://github.com/huggingface/safetensors ) to permit the latest version.
- [Release notes](https://github.com/huggingface/safetensors/releases )
- [Changelog](https://github.com/huggingface/safetensors/blob/main/RELEASE.md )
- [Commits](https://github.com/huggingface/safetensors/compare/v0.3.1...v0.3.3 )
---
updated-dependencies:
- dependency-name: safetensors
dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com >
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-01-15 22:34:40 +01:00
ea36f3b11f
Use the new phi model by default. ( #1589 )
2024-01-15 12:30:27 +01:00
79478ff5a1
Seed should be updated by random kernel result.
2024-01-15 11:58:25 +01:00
86b7c01b30
Update gemm to the latest version. ( #1587 )
2024-01-15 09:44:51 +01:00
bdd8107fda
Expose the ndarray trait. ( #1586 )
2024-01-14 20:09:49 +01:00
ecf88a6d38
Merge branch 'main' into ivarflakstad/metal-prng
2024-01-14 17:10:54 +01:00
e6d86b0819
Add the pow operator. ( #1583 )
...
* Add the pow operator.
* Support the pow operation in onnx.
2024-01-13 20:24:06 +01:00
88618255cb
Fix the rotary embeddings for the new phi implementation. ( #1582 )
...
* Fix the rotary embeddings for the new phi implementation.
* Match the activation.
* KV cache fix.
* Use the config activation function.
2024-01-13 19:44:41 +01:00
539ead927a
Update the Phi model to use the updated architecture. ( #1580 )
...
* Update the Phi model to use the updated architecture.
* Add more of the phi model.
* Repeat KV + caching.
* Apply the rotary embeddings.
* Add support for the new phi model in the phi example.
* Fix a couple glitches.
* Fix a couple more glitches.
2024-01-13 17:38:27 +01:00
a46864bd56
Fix "Minimal Mamba" link in README. ( #1577 )
2024-01-12 17:47:07 +01:00
bafe95b660
Fix format. ( #1576 )
2024-01-12 14:23:17 +01:00
a3d92ab226
Metal: Activate bfloat affine and add benchmark ( #1543 )
...
* Use cfg to seperate benchmark results based on features
* Add bfloat affine and benchmarks
* Fix flops calculation
* Remove allow pragma
* Avoid some unnecessary returns.
* Improve benchmarks layout
---------
Co-authored-by: Laurent <laurent.mazare@gmail.com >
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com >
2024-01-12 11:19:49 +01:00
e90bcdcc7c
Metal: f16 and bf16 where_cond + benchmark ( #1545 )
...
* Use cfg to seperate benchmark results based on features
* Add metal where_cond for f16 and bf16. Add benchmark
* Remove allow pragma
* Avoid some unnecessary returns.
* Improve benchmarks layout
* Updated feature separated benchmarks
---------
Co-authored-by: Laurent <laurent.mazare@gmail.com >
2024-01-12 11:18:11 +01:00
8e06bfb4fd
Mention VGG in the readme. ( #1573 )
2024-01-12 09:59:29 +01:00
6242276c09
Pin the revision used for phi-v2 + make it the default. ( #1572 )
...
* Pin the revision used for phi-v2 + make it the default.
* Tweak the custom-ops build.
2024-01-12 09:19:30 +01:00
e06e8d0dbe
fmt
2024-01-12 07:26:42 +01:00
e63bb8661b
Merge branch 'main' into ivarflakstad/metal-prng
2024-01-12 07:19:58 +01:00
41915184bb
Bugfix for dequantizing q5k layers. ( #1569 )
2024-01-11 23:15:11 +01:00
c1876b8041
Merge pull request #1567 from bayedieng/close-ifdef
2024-01-11 22:14:38 +01:00
85e5680277
remove metal version check
2024-01-11 21:02:03 +00:00
1327419776
close ifdef
2024-01-11 17:14:12 +00:00
402349d120
feat(bf16): add cast support + tests for cast + bin ops ( #1524 )
2024-01-11 15:49:13 +01:00
9f0c99f0c1
Seperate benchmarks by enabled features ( #1538 )
...
* Use cfg to seperate benchmark results based on features
* Remove allow pragma
* Avoid some unnecessary returns.
* Improve benchmarks layout
* Derive bench_name from actual device
* Run CPU benchmarks even when GPU feature is enabled
---------
Co-authored-by: Laurent <laurent.mazare@gmail.com >
2024-01-11 15:35:38 +01:00
0fc95c9f0c
Add a dequantize command to tensor-tools. ( #1565 )
...
* Add a dequantize command to tensor-tools.
* Clippy fixes.
2024-01-11 11:21:01 +01:00
