Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Fix unit tests for large/small tensor constants | Lester Solbakken | 2018-02-05 | 2 | -4/+5 |
| | |||||
* | Store small constants separately | Lester Solbakken | 2018-02-05 | 1 | -4/+11 |
| | |||||
* | Support small constants | Jon Bratseth | 2018-02-05 | 1 | -4/+17 |
| | |||||
* | Add logging when adding a constant | Jon Bratseth | 2018-02-05 | 1 | -1/+1 |
| | |||||
* | Merge pull request #4913 from vespa-engine/bratseth/feature-names-cleanup | Lester Solbakken | 2018-02-05 | 5 | -138/+13 |
|\ | | | | | Bratseth/feature names cleanup | ||||
| * | Unify canonicalization | Jon Bratseth | 2018-02-03 | 5 | -183/+13 |
| | | |||||
| * | Fetaure names cleanup WIP | Jon Bratseth | 2018-02-02 | 2 | -15/+60 |
| | | |||||
* | | No need to compress just to uncompress. | Henning Baldersheim | 2018-02-05 | 1 | -1/+9 |
|/ | |||||
* | Merge pull request #4859 from vespa-engine/balder/shortcut-on-timeout | Henning Baldersheim | 2018-02-02 | 28 | -132/+89 |
|\ | | | | | Balder/shortcut on timeout | ||||
| * | Nested namespaces | Henning Baldersheim | 2018-02-02 | 27 | -131/+88 |
| | | |||||
| * | Correctly test for expiry. | Henning Baldersheim | 2018-02-01 | 1 | -1/+1 |
| | | |||||
* | | Generalize dimension-wise | Jon Bratseth | 2018-02-02 | 2 | -7/+8 |
| | | |||||
* | | Replace / in Tensorflow constants and placeholders to _ | Lester Solbakken | 2018-02-02 | 5 | -14/+21 |
| | | |||||
* | | Allow type generalizations in if | Jon Bratseth | 2018-02-01 | 4 | -6/+75 |
| | | |||||
* | | Correct copyright headers | Jon Bratseth | 2018-02-01 | 2 | -2/+4 |
| | | |||||
* | | Canonicalize features | Jon Bratseth | 2018-01-31 | 6 | -13/+138 |
|/ | | | | | This allows us to find the type of features referenced in ranking expressions regardless of the form they are written in. | ||||
* | Nonfunctional changes only | Jon Bratseth | 2018-01-31 | 1 | -0/+1 |
| | |||||
* | Verify macros | Jon Bratseth | 2018-01-31 | 32 | -135/+162 |
| | |||||
* | Propagate type information through ranking expressions | Jon Bratseth | 2018-01-29 | 19 | -37/+141 |
| | |||||
* | Propagate type information | Jon Bratseth | 2018-01-29 | 11 | -33/+117 |
| | |||||
* | Don't use uint16_t when UChar is needed. | Tor Egge | 2018-01-27 | 3 | -2/+3 |
| | | | | Include needed unicode header file. | ||||
* | Merge pull request #4789 from ↵ | Henning Baldersheim | 2018-01-26 | 11 | -69/+138 |
|\ | | | | | | | | | vespa-engine/geirst/improve-alloc-strategy-in-data-store Geirst/improve alloc strategy in data store | ||||
| * | Use definition of huge page size from memory allocator. | Geir Storli | 2018-01-26 | 1 | -2/+1 |
| | | |||||
| * | Improve buffer allocation strategy in data store by matching underlying ↵ | Geir Storli | 2018-01-26 | 7 | -42/+95 |
| | | | | | | | | | | | | | | | | allocators. This should reduce the amount of memory wasted in allocations. 1) heap allocation: buffer size is power of 2 to match vespamalloc. 2) mmap allocation: buffer size is multiple of huge page size (2MB) to match mmap allocator. | ||||
| * | Refactor calculation of buffer allocation. | Geir Storli | 2018-01-25 | 4 | -27/+44 |
| | | |||||
* | | Refactor tensorflow import and add dropout test | Lester Solbakken | 2018-01-26 | 11 | -342/+3602 |
| | | |||||
* | | Correct copyright header | Jon Bratseth | 2018-01-25 | 1 | -1/+0 |
| | | |||||
* | | Add copyright headers | Jon Bratseth | 2018-01-25 | 4 | -0/+4 |
| | | |||||
* | | Remove logging | Jon Bratseth | 2018-01-25 | 2 | -1/+5 |
| | | |||||
* | | Refactor: Rename | Jon Bratseth | 2018-01-25 | 3 | -25/+24 |
| | | |||||
* | | Refactor: Move state to helper | Jon Bratseth | 2018-01-25 | 3 | -32/+29 |
| | | |||||
* | | Refactor: Extract test helper logic | Jon Bratseth | 2018-01-25 | 4 | -125/+166 |
| | | |||||
* | | Minor improvements | Jon Bratseth | 2018-01-25 | 1 | -12/+11 |
|/ | |||||
* | Keep imported tensor attribute vector read guard live during access. | Tor Egge | 2018-01-23 | 1 | -1/+2 |
| | |||||
* | Revert "Revert "Add imported tensor attribute vector"" | Tor Egge | 2018-01-23 | 12 | -2/+415 |
| | |||||
* | Revert "Add imported tensor attribute vector" | Håkon Hallingstad | 2018-01-23 | 12 | -415/+2 |
| | |||||
* | Merge pull request #4733 from ↵ | Tor Brede Vekterli | 2018-01-23 | 12 | -2/+415 |
|\ | | | | | | | | | vespa-engine/toregge/add-imported-tensor-attribute-vector Add imported tensor attribute vector | ||||
| * | Use ASSERT_TRUE instead of EXPECT_TRUE. | Tor Egge | 2018-01-23 | 1 | -1/+1 |
| | | | | | | | | Continuing on failure would lead to test program crashing. | ||||
| * | Fix typo in tensor spec. | Tor Egge | 2018-01-22 | 1 | -1/+1 |
| | | |||||
| * | Add class comment. | Tor Egge | 2018-01-22 | 1 | -0/+4 |
| | | |||||
| * | Add imported tensor attribute vector read guard. | Tor Egge | 2018-01-22 | 9 | -1/+217 |
| | | | | | | | | | | Extend unit tests for imported attribute vector to check imported tensor attribute. | ||||
| * | Add imported tensor attribute vector. | Tor Egge | 2018-01-22 | 7 | -1/+194 |
| | | |||||
* | | Merge branch 'master' into lesters/add-batch-norm-tf-case | Lester Solbakken | 2018-01-22 | 23 | -105/+267 |
|\ \ | |||||
| * | | Use double quotes | Jon Bratseth | 2018-01-22 | 2 | -4/+4 |
| | | | | | | | | | | | | Single quotes are not supported in the backend. | ||||
| * | | Merge pull request #4734 from ↵ | Henning Baldersheim | 2018-01-22 | 21 | -101/+262 |
| |\ \ | | |/ | |/| | | | | | | | vespa-engine/geirst/more-conservative-buffer-allocation-strategy-in-datastore-buffers Geirst/more conservative buffer allocation strategy in datastore buffers | ||||
| | * | Add config for the alloc grow factor used when allocating buffers in ↵ | Geir Storli | 2018-01-22 | 3 | -9/+8 |
| | | | | | | | | | | | | multi-value attributes. | ||||
| | * | Add more conservative buffer allocation strategy in datastore buffers. | Geir Storli | 2018-01-19 | 19 | -78/+243 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This should greatly reduce the memory footprint of multi-value attribute vectors which uses an array store for underlying values. The size of a new buffer is calculated as (simplified): *) before: (size of previous buffer) * 2 *) now: (size of all previous buffers) * allocation grow factor (default = 0.2) In addition we now support resizing of all buffers, not only the first one. | ||||
| | * | Style changes. | Geir Storli | 2018-01-18 | 2 | -16/+13 |
| | | | |||||
* | | | Support negative values dimension sizes in tensorflow reshape import | Lester Solbakken | 2018-01-22 | 1 | -2/+4 |
| | | | |||||
* | | | Support negative dimesions in import of expanddims | Lester Solbakken | 2018-01-22 | 1 | -3/+7 |
| | | |