ativilambit/neu-faq

README.md

@@ -1,6 +1,6 @@
 ---
 license: apache-2.0
-base_model: google/flan-t5-small
+base_model: google/flan-t5-base
 tags:
 - generated_from_trainer
 metrics:
@@ -15,13 +15,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # results
 
-This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on an unknown dataset.
+This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.6818
-- Rouge1: 0.1641
-- Rouge2: 0.0847
-- Rougel: 0.1425
-- Rougelsum: 0.1553
+- Loss: 1.0912
+- Rouge1: 0.2091
+- Rouge2: 0.1205
+- Rougel: 0.1845
+- Rougelsum: 0.1965
 
 ## Model description
 
@@ -46,14 +46,15 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 2
+- num_epochs: 3
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
 |:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|
-| No log        | 1.0   | 17   | 1.8080          | 0.1584 | 0.0740 | 0.1386 | 0.1486    |
-| No log        | 2.0   | 34   | 1.6818          | 0.1641 | 0.0847 | 0.1425 | 0.1553    |
+| No log        | 1.0   | 17   | 1.3457          | 0.2023 | 0.0986 | 0.1645 | 0.1834    |
+| No log        | 2.0   | 34   | 1.1429          | 0.2047 | 0.1129 | 0.1777 | 0.1930    |
+| No log        | 3.0   | 51   | 1.0912          | 0.2091 | 0.1205 | 0.1845 | 0.1965    |
 
 
 ### Framework versions

config.json

@@ -1,12 +1,12 @@
 {
-  "_name_or_path": "google/flan-t5-small",
+  "_name_or_path": "google/flan-t5-base",
   "architectures": [
     "T5ForConditionalGeneration"
   ],
   "classifier_dropout": 0.0,
-  "d_ff": 1024,
+  "d_ff": 2048,
   "d_kv": 64,
-  "d_model": 512,
+  "d_model": 768,
   "decoder_start_token_id": 0,
   "dense_act_fn": "gelu_new",
   "dropout_rate": 0.1,
@@ -18,9 +18,9 @@
   "layer_norm_epsilon": 1e-06,
   "model_type": "t5",
   "n_positions": 512,
-  "num_decoder_layers": 8,
-  "num_heads": 6,
-  "num_layers": 8,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
   "output_past": true,
   "pad_token_id": 0,
   "relative_attention_max_distance": 128,

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d92e5c765ca680975fe3b4fbf8e5f63b8c703900df8e959d114bf14f5dac7880
-size 307907461
+oid sha256:476f316577859fa05ada88d46232bf4cec0d162779e92e1263c940fd3da57ba7
+size 990404917

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:31ce122897f6107898af6f06686840f883b3f952bbbef9fed20b8b59be30848b
+oid sha256:915193a4b9056fbf63b9cf04c29f8c200134da2c688e2d3d7ce4b762bdb4841a
 size 4155