Remove kOnlyAllocate - no longer used. Rename ReadOrAllocate -> ReadFromBlobs.
Rename Reshape -> Fixup to reflect the new scope.
Remove no longer used ShrinkRows.
This simplifies gemma-inl and is a prerequisite for removing ConstMat
(whose .ofs was previously used for merged tensors)
PiperOrigin-RevId: 758214083
KVCache: add ctor required for MatStorageT, remove Create; bf_pre_ffw_rms_out -> pre_ffw_rms_out
optimize_test: larger vocab_size requires more steps
shared.h: Remove unused u128 type
correctly set Activation matrix rows, avoid passing as arg
ops: pass Mat instead of pointers/sizes; vectorize LayerNorm; support any weight type
mat: add OverrideRows, used by SetBatchSize
PiperOrigin-RevId: 757790736
Also in this change:
* The computation for a single `q` is factored out and exposed.
* Strided `ConstMat` views into the KV caches are introduced to enable experimentation with various KV cache layouts.
PiperOrigin-RevId: 756339313
- chore: unused parameters cleaned up
- bugfix: explicitly use hwy::Span in GenerateInternal() to prevent runtime crashes due to memory layout incompatibility
- bugfix: explicit nullptr check in LogDebug
- chore: length-related parameters renamed for clarity
- feature: SaveConversation() can be optionally used to save copy of a conversation that ResetConversation() will rewind to upon request, rather than just an empty KV cache
- feature: GetCurrentConversation() can be used to query the current conversation's name
PiperOrigin-RevId: 755873147
Weight handling:
- new ModelStore2 supports both pre-2025 multi-file and single-file formats
- simpler ForEachTensor with TensorArgs
- tensors are constructed with their full suffixed name
I/O:
- support mmap and stride
- Simplified SbsWriter, single insert(); add SbsReader
Misc:
- kMockTokenizer: allow creating with unavailable tokenizer
- configs.h: Simpler enum validity checks via kSentinel
- matmul.h: remove unused enable_bind (now in allocator.h)
- tensor_info: single TensorInfoRegistry class, rename from tensor_index.h
Frontends:
- Replace Allocate/CreateGemma with ctor(LoaderArgs, MatMulEnv&)
- Deduce model/weight type, remove --model and parsing
- Replace most common.h includes with configs.h
- Remove --compressed_weights, use --weights instead
- Remove ModelInfo, replaced by ModelConfig.
Backprop:
- Reduce max loss, remove backward_scalar_test (timeout)
- Update thresholds because new RandInit changes rng eval order and thus numerics
PiperOrigin-RevId: 755317484
This change adds a basic C API that allows access to Gemma functionality from other programming languages. The functionality is exposed via a shared library (DLL on Windows), with C++ interfaces and a basic C# interop wrapper included.
To build the DLL, use the `windows-dll` preset, which includes the C and C++ sources as follows:
```
cmake --preset windows-dll
cmake --build --config Release --preset windows-dll -j 4
```
This should generate a `gemma.dll` in `<build-dir>/Release`.
To build for non-Windows, the appropriate C++ DLL linking will need to be done to generate a shared library for the target OS.
PiperOrigin-RevId: 750246272
- Added prompt flag to InferenceArgs for non-interactive mode
- Set user-facing options to verbosity level 1
- Fixed prompt_size declaration and variable ordering in run.cc
- Properly set prompt_size after WrapAndTokenize calls
- Moved kVerboseLogTokens block after prompt_size is set
This change adds a --prompt command-line option that allows users to
provide prompts directly without entering interactive mode, which is
useful for scripting and automation.
use new ThreadingContext2 instead of monostate/init in each frontend
Add ThreadingArgs(replaces AppArgs)
backprop: use Packed() accessor and MakePacked factory and row-based access to allow for stride
compress_weights: remove, moving to py-only exporter instead
Move MatPtr to mat.h and revise interface:
- Generic MatOwner
- rename accessors to Packed*
- support stride/row accessors, fix RowPtr stride
Add TypeBits(Type)
Move GenerateMat to test_util-inl for sharing between matmul test/bench
Move internal init to gemma.cc to avoid duplication
Rename GemmaEnv model_ to gemma_ for disambiguating vs upcoming ModelStorage
Remove --compressed_weights, use --weights instead.
tensor_index: add ExtentsFromInfo and TensorIndexLLM/Img
Allocator: use normal unique_ptr for AllocBytes so users can call directly
threading: use -> because AlignedPtr no longer assumes arrays
PiperOrigin-RevId: 745918637
Gemma receives a MatMulEnv arg, with comment on lifetime
Split threading into topology so the latter can be used in allocator
Add AllocClasses() for non-POD (ThreadPool)
Support binding pool to NUMA node
Update threading_test with latency measurements
Also update Highway version.
PiperOrigin-RevId: 736904748
Adds Stride to ConstMat, to support decompression of C output for test
matmul_test: add line numbers to output
Also ignore "N is not a multiple of nc" when N==nc
PiperOrigin-RevId: 731096662
This computes softmax on the top-K logits, instead of computing softmax first and then getting top-K probs. So we end up avoiding renormalizing too. Additionally, modify softmax to do temperature scaling, if temp != 1.0
PiperOrigin-RevId: 727702149
This will be useful for adding sampling functionality like beam decoding, parallel sampling, cot decoding (as described in the [Chain-of-Thought Reasoning Without Prompting paper](https://arxiv.org/abs/2402.10200))
PiperOrigin-RevId: 725151530
ops.h: move CreateInvTimescale to allow calling without depending on gemma
Pass around MatMulEnv instead of pools to avoid re-creating the env
profiler.h can now be used outside SIMD code
allocator: add StepBytes and QuantumSteps
rename worker thread with package/cluster in the name
threading: add Visit* to IndexRange
PiperOrigin-RevId: 718766704
Use `std::size` for array size calculations. Replace C-style
string manipulations with `std::string` methods. Simplify
`std::transform` usage for case conversion.
Signed-off-by: Eric Curtin <ecurtin@redhat.com>
- use HalfRope position encodings
- zero-initialize the caches for each Generate at position 0
The lack of the latter made the tests in gemma_test dependent on each other.
PiperOrigin-RevId: 694509054
Add Extents2D, Range2D vocab types
Matmul uses ConstMat for inputs and RowPtr for output
Move RowVectorBatch to basics.h
Separate threading.cc
Fix topology string: report cores not LPs, and #HT
Move QStride/IsMHA into LayerConfig
ImageTokens does not require make_unique.
matmul_test: no longer require template args
PiperOrigin-RevId: 692963605
Use ModelConfig values for ImageTokens.
Output timing info for image token generation.
Add a method to copy image data into Image class directly.
Minor changes: pipe ModelTraining to more places.
PiperOrigin-RevId: 690572283
Improved threading.h, fix thread counts for single package/cluster systems
Temporarily forces to a single socket. Prefill 29.28 tps, decode 6.92.
Also fix benchmarks.cc build, update tensor allocator to Allocator
PiperOrigin-RevId: 687307167
Move image related config values from LayerConfig to ModelConfig.
Minor changes: Add a few comments, remove gcpp:: qualification where it wasn't needed in a few places, define local constants in VitAttention.DotSoftmaxWeightedSum()
PiperOrigin-RevId: 687210519
Changed CompressedLayer and CompressedWeights to be constructed with an instance of a LayerConfig and WeightsConfig respectively.
Added CompressedModel to remove ByteStorageT and get rid of most of the type casting, as well as allowing the default destructor to be used and work properly.
Adjusted WeightsWrapper and ForwardLayer etc to match.
The only remaining template arg is the weight type.
This enables all the instantiations to be deleted, apart from one per type.
It also enables (but not yet done) the config to be stored in the blob file instead of having to be specified separately.
Reduces the size of the gemma_lib and weights shared libraries by a factor of 4.3 and 3.2 respectively.
PiperOrigin-RevId: 686870060
add app.h comment
compress-inl: remove unused typedef
gemma-inl: add missing HWY_ATTR and cast
separate sum-inl.h and basics.h headers
replace more hwy::bfloat16_t with BF16
update include pragmas
update dot_test thresholds
update Highway version in Bazel for HWY_RCAST_ALIGNED fix
PiperOrigin-RevId: 684464326
Definition of array size is moved to the constructor.
Allocation is separate and parallelized.
All users of weights_raw.h migrated to CompressedWeights and weights_raw.h deleted.
Replaced all previous ForEachTensor functions with a single unified function.
PiperOrigin-RevId: 684451604
Use in dot_test
app.h: add new flags and rename num_threads to max_threads
matmul: Parallelize MatMulSlow and enable spinning, more large/fewer medium test cases
PiperOrigin-RevId: 683216386
See https://arxiv.org/abs/2407.07726 for a description of the model.
Because PaliGemma operates as a prefix-LM on the image+prompt, add support for that.
PiperOrigin-RevId: 677841119
Compression:
* Implement {any packed} x {bf16, f32} 'Load2' and DecompressAndZeroPad
* New compression test for all packed formats, add to GEMMA_TEST_FILES, remove from sfp/nuq_test
* Decompress->DecompressAndZeroPad, use PackedSpan for args with bounds checking
* NUQ: support arbitrary-length enc/dec
* New compression/shared, remove sfp.h and nuq.h
* Move Store2 into Traits and provide Compress2 wrapper
* Remove unused Decompress()-with-pool overload
* Simplify CompressedArrayLen, rename to CompressedArrayElements
* Remove unused DistortionStats b_l1_
Misc:
* Add compensated and Kahan dot, support any length
* Use same Dot function everywhere
* Move exact arithmetic functions into fp_arith
* use FloatPtr and MatPtr typedefs in tests; less stack usage
* Rename args to packed/raw
* Remove Traits::Name, instead TypeName<T>()
* Move kMaxSFP and kClusters/kGroupSize into Sfp/NuqStream
PiperOrigin-RevId: 672868468
Use args only in GemmaEnv constructor, store everything else in RuntimeConfig.
Add runtime option to turn off thread spinning.
PiperOrigin-RevId: 670467320
Supports converting all weight/activation formats to native MulT (bf16/f32)
Also:
- ConstMat/MutableMat for const correctness
- Move RowVectorBatch to allocator.h so it can be used from Matmul
- Add matmul.h so MatMulEnv can be used from Activations
- Remove kMaxThreads, detect from PerClusterPools
- Build fix: -inl.h files must be textual_hdrs, and highway.h should precede -inl.h
```
zen4 new
64, 24576, 3072, add=0, MatTA=bf16, MatTB=sfp: 616.6 GFLOPS.
64, 3072, 24576, add=0, MatTA=bf16, MatTB=sfp: 460.7 GFLOPS.
64, 24576, 3072, add=0, MatTA=f32, MatTB=sfp: 598.6 GFLOPS.
64, 3072, 24576, add=0, MatTA=f32, MatTB=sfp: 435.6 GFLOPS.
zen4 old
64, 24576, 3072, add=0, MatTA=f32, MatTB=sfp: 257.5 GFLOPS.
64, 3072, 24576, add=0, MatTA=f32, MatTB=sfp: 231.9 GFLOPS.
```
PiperOrigin-RevId: 663729812
LayersOutputFunc is no longer invoked for "blocks" and "final_norm" outputs.
Instead, we directly expose the Activations structure.
PiperOrigin-RevId: 663409316
Split attention into functions, move into class.
Fuse Rope and MulBy, allow non-in-place version to avoid copy from q to KV.
Sink if() into MaybeLogitsSoftCap.
PiperOrigin-RevId: 661168418
Limit thread counts to detected. Add max_clusters arg.
Update detection logic to check for smt0 - previously we pinned to some siblings.
PiperOrigin-RevId: 659755311
- Allocate only the required KV caches and activation batch size
- Add flags for batch sizes
- Const-correct interface: Span of const int.
- Also clean up the KVCache arg to a span.
- Move kPrefillBatchSize into RuntimeConfig and remove related global constants.
PiperOrigin-RevId: 655893197
SfpCodec::Dec2F and ComressTraits<T>::Decompress2 for all supported types. It also allows to remove one of the specializations of GEMM_4x4_Tile, handling compressed MatB with one function. As before even when MatA is bf16 it is using 32-bit registers for computations.
Measurements for a 2b-it sfp-encoded model on a AMD Ryzen Threadripper PRO 3945WX 12-Cores:
baseline:
```
32.6254 prefill tokens / sec
8.91429 tokens / sec
115 milliseconds time to first token
```
this change:
```
54.3045 prefill tokens / sec
16.8191 tokens / sec
56 milliseconds time to first token
```
PiperOrigin-RevId: 651369694
This CL adds a new function to Gemma that allows for batching of multiple prompts. The function takes a vector of prompts and returns a vector of responses. The prompts are processed in parallel, and the responses are returned in the same order as the prompts.
PiperOrigin-RevId: 648367559
For regular (non-SSM) Gemma models, kGriffinLayers is by definition always zero
and kGemmaLayers is just the number of layers.
PiperOrigin-RevId: 644384531
Helper functions to tokenize/wrap
Move LayersOutputFunc into RuntimeConfig
AcceptFunc passes the probability
Implement StringFromType using the parser, and verify results match
PiperOrigin-RevId: 643255119
accept_token: allow default, check if empty when using
allow mixing sample_func and stream_func, call the latter after the former
Also fix missing includes/deps.
PiperOrigin-RevId: 642240012
Jan Wassenberg
RangerUFO
Biruk Mammo
The gemma.cpp Authors
prajwalc22
Prajwal Choudhari
Copybara-Service
Phil Culliton
Apoorv Reddy
Daniel Keysers
Ray Smith
Eric Curtin
Krzysztof Ostrowski
Zoltan Szabadka
Paul Chang
Andrey Vlasov
Kan Wu