From ec543a8a8388b5356419b24ce3d77acaefb07896 Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Sat, 25 Nov 2017 14:21:59 -0500
Subject: [PATCH] update models_class1_experiments1 to include embedding and
 onehot input variations

---
 .../models_class1/hyperparameters.yaml        |  1 +
 .../models_class1_experiments1/GENERATE.sh    |  4 +-
 .../hyperparameters-embedding.yaml            | 50 +++++++++++++++++++
 .../hyperparameters-onehot.yaml               | 49 ++++++++++++++++++
 4 files changed, 102 insertions(+), 2 deletions(-)
 create mode 100644 downloads-generation/models_class1_experiments1/hyperparameters-embedding.yaml
 create mode 100644 downloads-generation/models_class1_experiments1/hyperparameters-onehot.yaml

diff --git a/downloads-generation/models_class1/hyperparameters.yaml b/downloads-generation/models_class1/hyperparameters.yaml
index 30b78fa2..7984d882 100644
--- a/downloads-generation/models_class1/hyperparameters.yaml
+++ b/downloads-generation/models_class1/hyperparameters.yaml
@@ -26,6 +26,7 @@
 # One of "one-hot", "embedding", or "BLOSUM62".
 "peptide_amino_acid_encoding": "BLOSUM62",
 "use_embedding": false,  # maintained for backward compatability
+"embedding_output_dim": 8,  # only used if using embedding
 "kmer_size": 15,
 
 ##########################################
diff --git a/downloads-generation/models_class1_experiments1/GENERATE.sh b/downloads-generation/models_class1_experiments1/GENERATE.sh
index 148bd15f..379a53f6 100755
--- a/downloads-generation/models_class1_experiments1/GENERATE.sh
+++ b/downloads-generation/models_class1_experiments1/GENERATE.sh
@@ -37,9 +37,9 @@ time mhcflurry-class1-train-allele-specific-models \
     --alleles $ALLELES &
 
 # Model variations on qualitative + quantitative
-for mod in 0local_noL1 0local 1local dense16 dense64 noL1 
+for mod in 0local_noL1 0local 2local dense16 dense64 noL1 onehot embedding
 do
-    cp $SCRIPT_DIR/hyperparameters-${mod}.json .
+    cp $SCRIPT_DIR/hyperparameters-${mod}.yaml .
     mkdir models-${mod}
     time mhcflurry-class1-train-allele-specific-models \
         --data "$(mhcflurry-downloads path data_curated)/curated_training_data.csv.bz2" \
diff --git a/downloads-generation/models_class1_experiments1/hyperparameters-embedding.yaml b/downloads-generation/models_class1_experiments1/hyperparameters-embedding.yaml
new file mode 100644
index 00000000..40062bd5
--- /dev/null
+++ b/downloads-generation/models_class1_experiments1/hyperparameters-embedding.yaml
@@ -0,0 +1,50 @@
+[{
+##########################################
+# ENSEMBLE SIZE
+##########################################
+"n_models": 8,
+
+##########################################
+# OPTIMIZATION
+##########################################
+"max_epochs": 500,
+"patience": 10,
+"early_stopping": true,
+"validation_split": 0.2,
+
+##########################################
+# RANDOM NEGATIVE PEPTIDES
+##########################################
+"random_negative_rate": 0.0,
+"random_negative_constant": 25,
+"random_negative_affinity_min": 20000.0,
+"random_negative_affinity_max": 50000.0,
+
+##########################################
+# PEPTIDE REPRESENTATION
+##########################################
+# One of "one-hot", "embedding", or "BLOSUM62".
+"peptide_amino_acid_encoding": "embedding",
+"use_embedding": true,  # maintained for backward compatability
+"embedding_output_dim": 8,  # only used if using embedding
+"kmer_size": 15,
+
+##########################################
+# NEURAL NETWORK ARCHITECTURE
+##########################################
+"locally_connected_layers": [
+    {
+        "filters": 8,
+        "activation": "tanh",
+        "kernel_size": 3
+    }
+],
+"activation": "relu",
+"output_activation": "sigmoid",
+"layer_sizes": [
+    32
+],
+"dense_layer_l1_regularization": 0.001,
+"batch_normalization": false,
+"dropout_probability": 0.0,
+}]
diff --git a/downloads-generation/models_class1_experiments1/hyperparameters-onehot.yaml b/downloads-generation/models_class1_experiments1/hyperparameters-onehot.yaml
new file mode 100644
index 00000000..1742093d
--- /dev/null
+++ b/downloads-generation/models_class1_experiments1/hyperparameters-onehot.yaml
@@ -0,0 +1,49 @@
+[{
+##########################################
+# ENSEMBLE SIZE
+##########################################
+"n_models": 8,
+
+##########################################
+# OPTIMIZATION
+##########################################
+"max_epochs": 500,
+"patience": 10,
+"early_stopping": true,
+"validation_split": 0.2,
+
+##########################################
+# RANDOM NEGATIVE PEPTIDES
+##########################################
+"random_negative_rate": 0.0,
+"random_negative_constant": 25,
+"random_negative_affinity_min": 20000.0,
+"random_negative_affinity_max": 50000.0,
+
+##########################################
+# PEPTIDE REPRESENTATION
+##########################################
+# One of "one-hot", "embedding", or "BLOSUM62".
+"peptide_amino_acid_encoding": "one-hot",
+"use_embedding": false,  # maintained for backward compatability
+"kmer_size": 15,
+
+##########################################
+# NEURAL NETWORK ARCHITECTURE
+##########################################
+"locally_connected_layers": [
+    {
+        "filters": 8,
+        "activation": "tanh",
+        "kernel_size": 3
+    }
+],
+"activation": "relu",
+"output_activation": "sigmoid",
+"layer_sizes": [
+    32
+],
+"dense_layer_l1_regularization": 0.001,
+"batch_normalization": false,
+"dropout_probability": 0.0,
+}]
-- 
GitLab