Upload TFEsmForSequenceClassification

README.md

@@ -1,8 +1,8 @@
 ---
 license: mit
+base_model: facebook/esm2_t12_35M_UR50D
 tags:
 - generated_from_keras_callback
-base_model: facebook/esm2_t12_35M_UR50D
 model-index:
 - name: esm2_t12_35M_UR50D-finetuned-AMP_Antibacteria-classification
   results: []
@@ -15,11 +15,11 @@ probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/esm2_t12_35M_UR50D](https://huggingface.co/facebook/esm2_t12_35M_UR50D) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Train Loss: 0.2932
-- Train Accuracy: 0.9052
-- Validation Loss: 0.2976
-- Validation Accuracy: 0.9034
-- Epoch: 1
+- Train Loss: 0.1176
+- Train Accuracy: 0.9553
+- Validation Loss: 0.5906
+- Validation Accuracy: 0.8286
+- Epoch: 0
 
 ## Model description
 
@@ -45,8 +45,7 @@ The following hyperparameters were used during training:
 
 | Train Loss | Train Accuracy | Validation Loss | Validation Accuracy | Epoch |
 |:----------:|:--------------:|:---------------:|:-------------------:|:-----:|
-| 0.2991     | 0.9052         | 0.2984          | 0.9034              | 0     |
-| 0.2932     | 0.9052         | 0.2976          | 0.9034              | 1     |
+| 0.1176     | 0.9553         | 0.5906          | 0.8286              | 0     |
 
 
 ### Framework versions

config.json

@@ -10,9 +10,17 @@
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.0,
   "hidden_size": 480,
+  "id2label": {
+    "0": "negative",
+    "1": "positive"
+  },
   "initializer_range": 0.02,
   "intermediate_size": 1920,
   "is_folding_model": false,
+  "label2id": {
+    "negative": 0,
+    "positive": 1
+  },
   "layer_norm_eps": 1e-05,
   "mask_token_id": 32,
   "max_position_embeddings": 1026,

tf_model.h5

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ce9aa35ec49feebbba0d2ea64b5292c317b1f9764eebf8481ff7d07fbd01d4c4
+oid sha256:b8c30e54c9e78ad6ff72712c23b3f0c1d1a34e52894b8c877e5ad36f207e8730
 size 136284336