initial commit

README.md

@@ -0,0 +1,6 @@
+# iSEEEK
+Generative pretraining from the rankings of top expressing genes. 
+It was trained on more than 20 million single-cell transcriptomes with a sequence length of 64.
+
+
+

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "bert-1536-1-16-2023Jan10/config.json",
+  "architectures": [
+    "BertForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "cls_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1536,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_norm_eps": 1e-12,
+  "mask_token_id": 4,
+  "max_position_embeddings": 1024,
+  "model_type": "bert",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 1,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "sep_token_id": 3,
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.0.dev0",
+  "type_vocab_size": 2,
+  "unk_token_id": 1,
+  "use_cache": true,
+  "vocab_size": 21051
+}

cp.sh

@@ -0,0 +1,3 @@
+cp /data/ai/gene_expression/bert/bert-1536-1-16-2023Jan10/pytorch_model.bin ./
+cp /data/ai/gene_expression/bert/bert-1536-1-16-2023Jan10/config.json ./
+cp /data/ai/gene_expression/bert/berttokenizer/* ./

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1d3ec5f69a7178fc41905f21dd418ffaa0999c14c7295030c667f7018777cce
+size 258535764

special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"model_max_length": 512, "pad_token": "[PAD]", "unk_token": "[UNK]", "cls_token": "[CLS]", "sep_token": "[SEP]", "mask_token": "[MASK]", "tokenizer_class": "PreTrainedTokenizerFast"}