Upload model

config.json

@@ -0,0 +1,57 @@
+{
+  "_name_or_path": "allenai/led-base-16384",
+  "activation_dropout": 0.0,
+  "activation_function": "gelu",
+  "architectures": [
+    "CustomLEDForResultsIdModel"
+  ],
+  "attention_dropout": 0.0,
+  "attention_window": [
+    1024,
+    1024,
+    1024,
+    1024,
+    1024,
+    1024
+  ],
+  "auto_map": {
+    "AutoModel": "modeling_CustomLEDForResultsId.CustomLEDForResultsIdModel"
+  },
+  "bos_token_id": 0,
+  "classif_dropout": 0.0,
+  "classifier_dropout": 0.0,
+  "d_model": 768,
+  "decoder_attention_heads": 12,
+  "decoder_ffn_dim": 3072,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 2,
+  "dropout": 0.1,
+  "encoder_attention_heads": 12,
+  "encoder_ffn_dim": 3072,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2"
+  },
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2
+  },
+  "max_decoder_position_embeddings": 1024,
+  "max_encoder_position_embeddings": 16384,
+  "model_type": "led",
+  "num_hidden_layers": 6,
+  "pad_token_id": 1,
+  "torch_dtype": "float32",
+  "transformers_version": "4.35.2",
+  "use_cache": true,
+  "vocab_size": 50265
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7d1adfe6f64aa4a7d99d95064349361812c3ececce79dbb12a8e4a005459b0f
+size 417408708

modeling_CustomLEDForResultsId.py

@@ -0,0 +1,44 @@
+# from transformers.models.led.modeling_led import LEDEncoder
+from transformers import LEDConfig, LEDModel, LEDPreTrainedModel
+from transformers.modeling_outputs import TokenClassifierOutput
+
+import torch.nn as nn
+
+class CustomLEDForResultsIdModel(LEDPreTrainedModel):
+    def __init__(self, config: LEDConfig, checkpoint): 
+        super().__init__(config)
+        self.num_labels = config.num_labels 
+        print("Configs")
+        print(config.num_labels)
+        print(config.dropout)
+
+        #Load Model with given checkpoint and extract its body
+        if (checkpoint):
+            self.led = LEDModel.from_pretrained(checkpoint, config=config).get_encoder()
+        else:
+            self.led = LEDModel(config).get_encoder()
+            
+        # self.model = LEDEncoder.from_pretrained(checkpoint, config=config)
+        self.dropout = nn.Dropout(config.dropout) 
+        self.classifier = nn.Linear(self.led.config.d_model,self.num_labels) # load and initialize weights
+
+    def forward(self, input_ids=None, attention_mask=None, labels=None, global_attention_mask=None, return_loss=True):
+        #Extract outputs from the body
+        outputs = self.led(input_ids=input_ids, attention_mask=attention_mask, global_attention_mask=global_attention_mask)
+        
+        sequence_output = self.dropout(outputs.last_hidden_state)
+        logits = self.classifier(sequence_output) # calculate losses [:,0,:].view(-1,768)      
+
+        # start_logits, end_logits = logits.split(1, dim=-1)
+        # start_logits = start_logits.squeeze(-1).contiguous()
+        # end_logits = end_logits.squeeze(-1).contiguous()
+        
+        loss = None
+        if labels is not None:
+            loss_fct = nn.CrossEntropyLoss()
+            loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
+
+        return {
+            'loss': loss,
+            'logits': logits
+        }