TypeError: tuple indices must be integers or slices, not str

Hi all,

I have resubmitted the same job as my [previous issue](https://github.com/deepmind/alphafold/issues/742), except this time on GeForce RTX 2080 Ti rather than A10 or A40s (under the assumption that these latter two need to have drivers updated). This gave a new error below.

thanks in advance!

```
I0406 08:08:54.468482 23367702087488 templates.py:256] Found an exact template match 2hxp_A.
I0406 08:08:54.631971 23367702087488 templates.py:256] Found an exact template match 4bvx_B.
I0406 08:08:54.644156 23367702087488 pipeline.py:234] Uniref90 MSA size: 36 sequences.
I0406 08:08:54.644257 23367702087488 pipeline.py:235] BFD MSA size: 36 sequences.
I0406 08:08:54.644309 23367702087488 pipeline.py:236] MGnify MSA size: 3 sequences.
I0406 08:08:54.644356 23367702087488 pipeline.py:237] Final (deduplicated) MSA size: 75 sequences.
I0406 08:08:54.644787 23367702087488 pipeline.py:239] Total number of templates (NB: this can include bad templates and is later filtered to top 4): 20.
I0406 08:08:54.674234 23367702087488 run_alphafold.py:216] Running model model_1_multimer_v3_pred_0 on test_multi
I0406 08:08:54.674847 23367702087488 model.py:138] Running predict with shape(feat) = {'aatype': (480,), 'residue_index': (480,), 'seq_length': (), 'msa': (512, 480), 'num_alignments': (), 'template_aatype': (4, 480), 'template_all_atom_mask': (4, 480, 37), 'template_all_atom_positions': (4, 480, 37, 3), 'asym_id': (480,), 'sym_id': (480,), 'entity_id': (480,), 'deletion_matrix': (512, 480), 'deletion_mean': (480,), 'all_atom_mask': (480, 37), 'all_atom_positions': (480, 37, 3), 'assembly_num_chains': (), 'entity_mask': (480,), 'num_templates': (), 'cluster_bias_mask': (512,), 'bert_mask': (512, 480), 'seq_mask': (480,), 'msa_mask': (512, 480)}
I0406 08:17:29.640328 23367702087488 model.py:201] Output shape was {'aligned_confidence_probs': (480, 480, 64), 'distogram': {'bin_edges': (63,), 'logits': (480, 480, 64)}, 'experimentally_resolved': {'logits': (480, 37)}, 'iptm': (), 'masked_msa': {'logits': (508, 480, 22)}, 'max_predicted_aligned_error': (), 'mean_plddt': (), 'plddt': (480,), 'predicted_aligned_error': (480, 480), 'predicted_lddt': {'logits': (480, 50)}, 'ptm': (), 'ranking_confidence': (), 'structure_module': {'final_atom_mask': (480, 37), 'final_atom_positions': (480, 37, 3)}, 'tol': ()}
I0406 08:17:29.641377 23367702087488 run_alphafold.py:228] Total JAX model model_1_multimer_v3_pred_0 on test_multi predict time (includes compilation time, see --benchmark): 515.0s
Traceback (most recent call last):
File "/mnt/nfs/clustersw/Debian/bullseye/cuda/11.2/alphafold/2.3.4/alphafold-2.3.4/run_alphafold.py", line 468, in
app.run(main)
File "/mnt/nfs/clustersw/Debian/bullseye/cuda/11.2/alphafold/2.3.4/lib/python3.9/site-packages/absl/app.py", line 312, in run
_run_main(main, args)
File "/mnt/nfs/clustersw/Debian/bullseye/cuda/11.2/alphafold/2.3.4/lib/python3.9/site-packages/absl/app.py", line 258, in _run_main
sys.exit(main(argv))
File "/mnt/nfs/clustersw/Debian/bullseye/cuda/11.2/alphafold/2.3.4/alphafold-2.3.4/run_alphafold.py", line 443, in main
predict_structure(
File "/mnt/nfs/clustersw/Debian/bullseye/cuda/11.2/alphafold/2.3.4/alphafold-2.3.4/run_alphafold.py", line 242, in predict_structure
plddt = prediction_result['plddt']
TypeError: tuple indices must be integers or slices, not str
```