0x1orz

how to an experiment start from a fixed set of hyper-parameter within search space?

1

**Describe the issue**: A set of hyper-parameter had nice worked on my model before hyper-parameter searching. The set is still best than all of hyper-parameter searched by nni. How to...

```bash mixcr analyze amplicon \ --threads 32 \ --species mmu --starting-material dna \ --5-end no-v-primers --3-end j-primers \ --adapters no-adapters \ --receptor-type bcr --contig-assembly --region-of-interest VDJRegion --only-productive \ --assemble "-OassemblingFeatures=[FR1,CDR1,FR2,CDR2,FR3,CDR3,FR4]"...

test data and tags in the full-length BCR analysis tutorial

2

mixcr is an awesome toolkit for BCR-seq analysis, I follow the tutorial https://github.com/milaboratory/mixcr/wiki/Analysis-of-full-length-BCR-data-with-unique-molecular-identifiers, with some confusions: 1. how to assess to the example files (`data_R1.fastq.gz` , `data_R2.fastq.gz` )? 2. what...

cell umi pattern

1

for the umi pattern `"^(CELL:N{12})(UMI:N{6})(R1:*)"` , it raise the following error in 4.0.0-68-developer with analyze pipeline while it works in 4.0.0 `align` subcommand : ```plainttext picocli.CommandLine$ExecutionException: Error while running command...

failed to correctAndSortTags

1

cleaned reads have two UMI and primers located the V(D)J , with patterns like `^(MIF:N{:6})tcga(R1:*) \ ^(MIR:N{:6})ctag(R2:*)` ```plantext java --version openjdk 17.0.3-internal 2022-04-19 OpenJDK Runtime Environment (build 17.0.3-internal+0-adhoc..src) OpenJDK 64-Bit...

How about biosignal?

1

Hay.Nowdays, I am interesting in Biomedical Signal Processing by deep learning. I am confused by the biosignal sequence to extract data for learning model.So that I had to study the...

how to get v_identity, j_identity against with germline ?

1

how to get v_identity, j_identity against with germline , of which a output of anarci if assign_germline ?

[Feature Request] Enable nn.Module with Python copy

6

To reuse a module copy, like: ```python from copy import deepcopy import mlx.nn as mlx_nn repeat = lambda module, n: [deepcopy(module) for _ in range(n)] mlx_layer = mlx_nn.Linear(2, 3) mlx_layers...

As a reset_parameters function of some layer usually is custom in PyTorch, while that is missing in MLX.

the requirement bio-embeddings-esm in meta.yaml is inconsistent with in readme `fair-esm`

1

`conda/torchdrug/meta.yaml` requires `bio-embeddings-esm `(latest 0.2.1) is outed-of-date, while in readme requires `fair-esm` (latest 2.0).