Merge branch 'v4-models' into 'master'

kit 14 v4.0.0 models

See merge request machine-learning/bonito!122

.gitlab-ci.yml

@@ -16,7 +16,7 @@ sre-test:
           "script_parameters": {
               "bonito_version":"'${CI_COMMIT_SHA}'",
               "sre_output":"'${SRE_OUTPUT_ROOT}'/'${CI_COMMIT_SHA}'/",
-              "models":["dna_r10.4.1_e8.2_400bps_fast@v3.5.2","dna_r10.4.1_e8.2_400bps_hac@v3.5.2","dna_r10.4.1_e8.2_400bps_sup@v3.5.2"],
+              "models":["dna_r10.4.1_e8.2_400bps_fast@v4.0.0","dna_r10.4.1_e8.2_400bps_hac@v4.0.0","dna_r10.4.1_e8.2_400bps_sup@v4.0.0"],
               "validation_set_path":"'${SRE_OUTPUT_ROOT}'/validation_sets",
               "validation_data_type":"prom"
           }

README.md

@@ -13,7 +13,7 @@ A PyTorch Basecaller for Oxford Nanopore Reads.
 ```bash
 $ pip install --upgrade pip
 $ pip install ont-bonito
-$ bonito basecaller dna_r10.4_e8.1_sup@v3.4 /data/reads > basecalls.bam
+$ bonito basecaller dna_r10.4_e8.1_sup@v4.0.0 /data/reads > basecalls.bam
 ```
 
 By default `pip` will install `torch` which is build against CUDA 10.2. For CUDA 11.3 builds run:
@@ -25,7 +25,7 @@ $ pip install --extra-index-url https://download.pytorch.org/whl/cu113 ont-bonit
 Bonito supports writing aligned/unaligned `{fastq, sam, bam, cram}`.
 
 ```bash
-$ bonito basecaller dna_r10.4_e8.1_sup@v3.4 --reference reference.mmi /data/reads > basecalls.bam
+$ bonito basecaller dna_r10.4_e8.1_sup@v4.0.0 --reference reference.mmi /data/reads > basecalls.bam
 ```
 
 Bonito will download and cache the basecalling model automatically on first use but all models can be downloaded with -
@@ -40,7 +40,7 @@ $ bonito download --models         # download all available models
 Modified base calling is handled by [Remora](https://github.com/nanoporetech/remora).
 
 ```bash
-$ bonito basecaller dna_r10.4_e8.1_sup@v3.4 /data/reads --modified-bases 5mC --reference ref.mmi > basecalls_with_mods.bam
+$ bonito basecaller dna_r10.4_e8.1_sup@v4.0.0 /data/reads --modified-bases 5mC --reference ref.mmi > basecalls_with_mods.bam
 ```
 
 To use the GPU-powered modified bases inference the `onnxruntime-gpu` package is required. 
@@ -59,7 +59,7 @@ $ bonito train --directory /data/training/ctc-data /data/training/model-dir
 In addition to training a new model from scratch you can also easily fine tune one of the pretrained models.  
 
 ```bash
-bonito train --epochs 1 --lr 5e-4 --pretrained dna_r10.4_e8.1_sup@v3.4 --directory /data/training/ctc-data /data/training/fine-tuned-model
+bonito train --epochs 1 --lr 5e-4 --pretrained dna_r10.4_e8.1_sup@v4.0.0 --directory /data/training/ctc-data /data/training/fine-tuned-model
 ```
 
 If you are interested in method development and don't have you own set of reads then a pre-prepared set is provide.

bonito/cli/download.py

@@ -78,6 +78,14 @@ def download(self):
 
 
 models = {
+    "[email protected]": "xsif8wbis1uyzsycopiql19p5mbd814n.zip",
+    "[email protected]": "cugqp8h9jrpfdgo696tzx5y3pxbjaqfl.zip",
+    "[email protected]": "fdxzet0grsp204a76mur3uiaeyx3i1x0.zip",
+
+    "[email protected]": "kf25wfgmmcda0457yx8u4f0t9vb2yozb.zip",
+    "[email protected]": "q5ecqe8c1u1pq4lrnyx9ypq070x1gbb8.zip",
+    "[email protected]": "ah2iqb80gfg7xncbwle7sgr2kq8l9437.zip",
+
     "[email protected]": "7662ke4ccxp2s9645b9mdp68c6i26ait.zip",
     "[email protected]": "k2p23nmw4k86lb5b4rlbapo8v3fidypn.zip",
     "[email protected]": "wibs7gs3uu0vyf9dkjtvuez1zb9kvcpf.zip",