Job instance 264503

Stderr output

<core_client_version>8.0.4</core_client_version>
<![CDATA[
<stderr_txt>
03:56:07 (17252): wrapper (7.7.26016): starting
03:56:10 (17252): wrapper: running ../../projects/axiom.heliex.net/axiom_worker_2.17_windows_x86_64.exe (wu.json result.json)
[CPU] Using NumPy
[Worker] Saved 5 gradient files to contribute folder (gradient sharing)
INFO: Loaded model from cache: expert_0_v2.0.bin (42580994 params)
INFO: Found 6 local files to train on.
INFO: Converted 2624704 bits from local files.
[Worker] Running 30 batches for expert 0...
[Worker] FATAL ERROR: MemoryError: Unable to allocate 48.0 MiB for an array with shape (64, 64, 3072) and data type float32
Traceback (most recent call last):
  File "worker_boinc_app.py", line 1699, in main
  File "worker_boinc_app.py", line 1322, in step
  File "simple_ml.py", line 1176, in backward
    return self.expert.backward_softmax_cross_entropy(x, y_idx, logits)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "simple_ml.py", line 649, in backward_softmax_cross_entropy
    self._backward_chunk(x, y_idx, logits, sc['ctx'], grads, B_total)
  File "simple_ml.py", line 725, in _backward_chunk
    d_a1 = d_z2 @ L['W2'].T  # [B, S, d_ff]
           ~~~~~^~~~~~~~~~~
numpy._core._exceptions._ArrayMemoryError: Unable to allocate 48.0 MiB for an array with shape (64, 64, 3072) and data type float32

04:15:23 (17252): axiom_worker_2.17_windows_x86_64.exe exited; CPU time 867.968750
04:15:23 (17252): called boinc_finish(0)

</stderr_txt>
]]>