update pytorch to 2.6

[Delaunay]

No description provided.

[Delaunay]

milabench install --base /home/mila/d/delaunap/scratch/shared/ --config config/standard.yaml --force
milabench prepare --base /home/mila/d/delaunap/scratch/shared/ --config config/standard.yaml
milabench sharedsetup --network /home/mila/d/delaunap/scratch/shared/ --local /tmp/workspace
milabench run --base /tmp/workspace/ --config /network/shared/setup/milabench/config/standard.yaml

[Delaunay]

=================
Benchmark results
=================

System
------
cpu:      Intel(R) Xeon(R) Gold 5418Y
n_cpu:    48
product:  NVIDIA L40S
n_gpu:    4
memory:   46068.0

Breakdown
---------
bench                    | fail |   n | ngpu |           perf |   sem% |   std% | peak_memory |          score | weight
brax                     |    0 |   1 |    4 |     1056707.04 |   0.0% |   0.3% |        1312 |     1056707.04 |   1.00
diffusion-gpus           |    1 |   1 |    4 |            nan |   nan% |   nan% |         nan |            nan |   1.00
diffusion-single         |    4 |   4 |    1 |            nan |   nan% |   nan% |         nan |            nan |   0.00
dimenet                  |    4 |   4 |    1 |            nan |   nan% |   nan% |         nan |            nan |   1.00
dinov2-giant-gpus        |    1 |   1 |    4 |            nan |   nan% |   nan% |       46002 |            nan |   1.00
dinov2-giant-single      |    4 |   4 |    1 |            nan |   nan% |   nan% |       45610 |            nan |   0.00
dqn                      |    0 |   4 |    1 | 24889893116.49 |   1.6% |  90.4% |        1098 | 99431155727.72 |   0.00
bf16                     |    0 |   4 |    1 |         199.35 |   0.3% |   5.8% |        1278 |         800.69 |   0.00
fp16                     |    0 |   4 |    1 |         188.20 |   0.2% |   3.2% |        1278 |         754.07 |   0.00
fp32                     |    0 |   4 |    1 |          43.18 |   0.1% |   1.5% |        1656 |         172.74 |   0.00
tf32                     |    0 |   4 |    1 |         114.99 |   0.2% |   2.9% |        1656 |         460.63 |   0.00
bert-fp16                |    0 |   4 |    1 |         204.84 |   1.0% |  11.5% |       15526 |         833.98 |   0.00
bert-fp32                |    0 |   4 |    1 |          63.57 |   0.5% |   5.1% |       20660 |         256.12 |   0.00
bert-tf32                |    0 |   4 |    1 |         111.54 |   0.7% |   7.5% |       20660 |         451.28 |   0.00
bert-tf32-fp16           |    0 |   4 |    1 |         205.87 |   1.0% |  11.4% |       15526 |         837.98 |   1.00
reformer                 |    0 |   4 |    1 |          29.65 |   0.3% |   4.2% |       12940 |         119.08 |   1.00
t5                       |    0 |   4 |    1 |          30.86 |   0.3% |   5.1% |       33876 |         124.14 |   0.00
whisper                  |    0 |   4 |    1 |         426.65 |   0.6% |   9.0% |        8752 |        1723.47 |   0.00
lightning                |    0 |   4 |    1 |         504.44 |   0.4% |   7.6% |       26098 |        2031.02 |   0.00
lightning-gpus           |    0 |   1 |    4 |        1992.85 |   0.0% |   0.2% |       26484 |        1992.85 |   1.00
llava-single             |    4 |   4 |    1 |            nan |   nan% |   nan% |       12222 |            nan |   1.00
llama                    |    0 |   4 |    1 |         294.77 |   6.8% |  86.9% |       27202 |        1117.16 |   1.00
llm-full-mp-gpus         |    0 |   1 |    4 |          31.81 |   3.1% |  16.6% |       25208 |          31.81 |   1.00
llm-lora-ddp-gpus        |    0 |   1 |    4 |        5243.89 |   0.4% |   2.3% |       32870 |        5243.89 |   1.00
llm-lora-mp-gpus         |    0 |   1 |    4 |         355.58 |   2.0% |  10.8% |       19428 |         355.58 |   1.00
llm-lora-single          |    0 |   4 |    1 |        2216.60 |   0.2% |   1.9% |       31112 |        8861.09 |   1.00
pna                      |    4 |   4 |    1 |            nan |   nan% |   nan% |         nan |            nan |   1.00
ppo                      |    0 |   4 |    1 |    62431321.23 |   0.7% |  58.0% |         980 |   249724763.51 |   1.00
recursiongfn             |    4 |   4 |    1 |            nan |   nan% |   nan% |         nan |            nan |   1.00
rlhf-gpus                |    0 |   1 |    4 |        6428.34 |   0.2% |   1.2% |       20428 |        6428.34 |   0.00
rlhf-single              |    0 |   4 |    1 |        1807.49 |   0.2% |   3.2% |       19326 |        7240.73 |   1.00
focalnet                 |    0 |   4 |    1 |         358.13 |   0.7% |  10.5% |       23034 |        1448.66 |   0.00
torchatari               |    0 |   4 |    1 |        7719.23 |   0.9% |  14.6% |        3274 |       30848.48 |   1.00
convnext_large-fp16      |    0 |   4 |    1 |         275.93 |   1.1% |  12.6% |       26866 |        1125.53 |   0.00
convnext_large-fp32      |    0 |   4 |    1 |          74.05 |   0.7% |   8.2% |       45498 |         299.94 |   0.00
convnext_large-tf32      |    0 |   4 |    1 |         104.05 |   0.8% |   9.4% |       45444 |         422.27 |   0.00
convnext_large-tf32-fp16 |    0 |   4 |    1 |         276.21 |   1.1% |  12.7% |       26866 |        1126.74 |   1.00
regnet_y_128gf           |    0 |   4 |    1 |          83.94 |   0.4% |   6.9% |       27504 |         338.20 |   1.00
resnet152-ddp-gpus       |    0 |   1 |    4 |        1999.47 |   0.0% |   0.2% |       26282 |        1999.47 |   0.00
resnet50                 |    0 |   4 |    1 |        1135.14 |   0.8% |  12.3% |       13298 |        4592.94 |   1.00
resnet50-noio            |    4 |   4 |    1 |            nan |   nan% |   nan% |         nan |            nan |   0.00
vjepa-gpus               |    1 |   1 |    4 |            nan |   nan% |   nan% |       36394 |            nan |   1.00
vjepa-single             |    4 |   4 |    1 |            nan |   nan% |   nan% |       45890 |            nan |   1.00

Scores
------
Failure rate:      24.65% (FAIL)
Score:             248.74

Errors
------
35 errors, details in HTML report.

[Delaunay]

torch geometric not yet updated for 2.6
https://pytorch-geometric.readthedocs.io/en/latest/install/installation.html