-
发版耗时由15min增加至40min
-
**Is your feature request related to a problem? Please describe:**
updates take longer than needed
**Describe the solution you'd like:**
using the prediction outputs, we can supply code and s…
-
runtime is ok, but given how fast ada-url is by itself, there is room to improvement at a) the interface R/C++ and b)the URLencoding to fix UTF8 support (see #1)
-
Once we have everything set up to run benchmarks, we should test if avoid cloning snapshots every time we do a far_call improves performance.
An initial idea is to use a [CoW](https://doc.rust-lang.o…
-
### Description
Running `MerkleHashBenchmarks` with
```
./gradlew :swirlds-unit-tests:common:swirlds-common-test:performanceTest --tests com.swirlds.common.test.merkle.MerkleHashBenchmarks
```
…
-
In /vllm/model_executor/weight_utils.py
```
elif use_safetensors:
for st_file in hf_weights_files:
with safe_open(st_file, framework="pt") as f:
for name in f.keys(): # noqa: SIM…
-
### Proposal to improve performance
We used the same GPU on two machines but different CPUs. The following experimental conclusions were drawn:
Experimental results: The GPU is 3090, and the CPU w…
-
```
diff --git a/tests/specs/mcd/concrete-rules.txt b/tests/specs/mcd/concrete-rules.txt
index c15aca708..8e7906129 100644
--- a/tests/specs/mcd/concrete-rules.txt
+++ b/tests/specs/mcd/concrete-r…
-
* see map2map distance functions in src/cryo_challenge/_map_to_map/map_to_map_distance.py
* add option to yaml for device CPU/GPU:0 etc.
* dist2dist (probably will not work on GPU)
* unsure if c…
-
It would be great to add some custom substructure filters to eliminate some undesired chemistry. NIBR filters built-in in the contrib dir is a good starting point, but it's single-threaded so it needs…