Modified validation and training for linktransformer model

.gitattributes

@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
+.git/lfs/objects/ac/e3/ace3c71b97a96f01b2feceec24958b690700ea37ad188af305b32ef472101546 filter=lfs diff=lfs merge=lfs -text

LT_training_config.json

@@ -1,10 +1,10 @@
 {
   "model_save_dir": "models",
   "model_save_name": "linkage_un_data_en_fine_fine",
-  "opt_model_description": "This model was trained on a dataset prepared by linking product classifications from [UN stats](https://unstats.un.org/unsd/classifications/Econ). \n                                 This model is designed to link different products together - trained on variation brought on by product level correspondance. It was trained for 30 epochs using other defaults that can be found in the repo's LinkTransformer config file - LT_training_config.json \n  ",
+  "opt_model_description": "This model was trained on a dataset prepared by linking product classifications from [UN stats](https://unstats.un.org/unsd/classifications/Econ). \n                                 This model is designed to link different products together - trained on variation brought on by product level correspondance. It was trained for 100 epochs using other defaults that can be found in the repo's LinkTransformer config file - LT_training_config.json \n  ",
   "opt_model_lang": "en",
   "train_batch_size": 64,
-  "num_epochs": 30,
+  "num_epochs": 100,
   "warm_up_perc": 1,
   "learning_rate": 2e-06,
   "val_perc": 0.2,

README.md

@@ -23,7 +23,7 @@ This model has been fine-tuned on the model : multi-qa-mpnet-base-dot-v1. It is
 
 
 This model was trained on a dataset prepared by linking product classifications from [UN stats](https://unstats.un.org/unsd/classifications/Econ). 
-                                 This model is designed to link different products together - trained on variation brought on by product level correspondance. It was trained for 30 epochs using other defaults that can be found in the repo's LinkTransformer config file - LT_training_config.json 
+                                 This model is designed to link different products together - trained on variation brought on by product level correspondance. It was trained for 100 epochs using other defaults that can be found in the repo's LinkTransformer config file - LT_training_config.json 
   
 
 ## Usage (LinkTransformer)
@@ -97,7 +97,7 @@ The model was trained with the parameters:
 
 **DataLoader**:
 
-`torch.utils.data.dataloader.DataLoader` of length 147 with parameters:
+`torch.utils.data.dataloader.DataLoader` of length 150 with parameters:
 ```
 {'batch_size': 64, 'sampler': 'torch.utils.data.dataloader._InfiniteConstantSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
 ```
@@ -109,8 +109,8 @@ The model was trained with the parameters:
 Parameters of the fit()-Method:
 ```
 {
-    "epochs": 30,
-    "evaluation_steps": 1470,
+    "epochs": 100,
+    "evaluation_steps": 1500,
     "evaluator": "sentence_transformers.evaluation.SequentialEvaluator.SequentialEvaluator",
     "max_grad_norm": 1,
     "optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
@@ -119,7 +119,7 @@ Parameters of the fit()-Method:
     },
     "scheduler": "WarmupLinear",
     "steps_per_epoch": null,
-    "warmup_steps": 4410,
+    "warmup_steps": 15000,
     "weight_decay": 0.01
 }
 ```

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ace3c71b97a96f01b2feceec24958b690700ea37ad188af305b32ef472101546
+oid sha256:a1caf77aeeb33f22a280e7dd98da99969ecbc0b510530aacd240d85af85b27f5
 size 438009257