update

app.py

@@ -64,8 +64,8 @@ with demo:
                 [![Downloads](https://pepy.tech/badge/pyabsa/month)](https://pepy.tech/project/pyabsa)
                 """
                 )
-    gr.Markdown("Your input text should be no more than 80 words, that's the longest text we used in training. However, you can try longer text in self-training ")
-    gr.Markdown("**You don't need to split each Chinese (Korean, etc.) token as the provided, just input the natural language text.**")
+    gr.Markdown("Your input text should be no more than 80 words, that's the longest text we used in training. However, you can train your own model using 512 max length")
+    gr.Markdown("**You don't need to split each Chinese (Korean, etc.) token as the provided examples, just input the natural language text.**")
     gr.Markdown("请确保输入的文本长度不超过200词，这是训练时的最大文本长度，过长将不会获得结果")
     gr.Markdown("**提供的中文等其他非拉丁语系数据集采用了空格分字，这是早期数据集的遗留问题，预测时不用对中文等语言进行空格分字**")
     output_dfs = []

checkpoint/English/TAD/TAD-Amazon.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:add4188835c415b72dda8f8db363bb82d6217b35f6b27a1dd9d82b029594607f
-size 427626189
+oid sha256:75427d8022884a26f385289601d22b877eda3f416b829d898f773e10f0d78cb6
+size 630751100

checkpoint/English/TAD/TAD-AmazonBAE.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4d11d8b01f3a5275906ab136739535adb423f34da0d99ff86d5e6a96a7035333
-size 618481827
+oid sha256:5c816bf76885d80fae3d8a17021e3d07339041cbcc0cb4fd24878353beb0f1ed
+size 618481549

checkpoint/English/TAD/TAD-AmazonPWWS.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e46273c6c696b230e9e344db28322bb692919c6de114cea8ced8de45ee713b86
-size 427631136
+oid sha256:68fca44dca3e18349b08d3c85529485c9fd43514f4e94146804db17959a0be8e
+size 621707291

checkpoint/English/TAD/TAD-AmazonTextFooler.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a4115d2d190f02cfd2cda0830ab013f011263d5b88129a0d997876f29ce6aa7
-size 628172415
+oid sha256:870fb9f96228793ef727bb71130b41c16b3e3b3e6adc616d42f95d85dbbfc1f6
+size 628174331

checkpoint/English/TAD/TAD-BERT-Amazon.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:383bf63ad88b690c894c4bbc29a8e9c0ac6175cea5e88bff1db197f06b8509c8
-size 427631665
+oid sha256:df1654efcfe101de7e42e134b55c4f37f7ffd37c6c1d5ef07295364e7b65337b
+size 427627904

checkpoint/English/TAD/TAD-BERT-AmazonBAE.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d700f72bc54f6931e83750d5d6850a77baf385effcb3b71b615b3009b3fd719a
-size 427631668
+oid sha256:1e8f2d39b40622153c4f7f8f3128e0a2aecd6fa71d59ec5cda22189d39b9d170
+size 427633731

checkpoint/English/TAD/TAD-BERT-AmazonPWWS.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:987a05b58a54ea533974be1fe943e7f069ec816eed23a4d57752950e705d7c63
-size 427631162
+oid sha256:40a51b5bf114f91e7909f6dc7d3d274e717ca4f38b7b2ff26a172ecf035bd4df
+size 427628535

checkpoint/English/TAD/TAD-BERT-AmazonTextFooler.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb8df1636dfa782a18ce09a7f4ececbcab83bbe68ae01f75bf0bed2e3b63d6ea
-size 427630967
+oid sha256:927b1cd28052c22377243a96f5d45a280cda44d878e83d9d9b4ca6381cbdd249
+size 427631361

release-note.json

@@ -1,4 +1,104 @@
 {
+  "1.16.6": {
+    "1": "Update findfile dependency to 1.7.9.7",
+    "2": ""
+  },
+  "1.16.5": {
+    "1": "Modify some output printing",
+    "2": "Fix offline pretrained model loading for text classification"
+  },
+  "1.16.4": {
+    "1": "Add validation set support for ATEPC",
+    "2": "Other improvements"
+  },
+  "1.16.3": {
+    "1": "fix patch for 1.16.2"
+  },
+  "1.16.2": {
+    "1": "fix a dataset detection issue",
+    "2": "Add probability distribution and confidence to the sentiment classification in ATEPC model",
+    "3": "Add a experimental feature to automatic annotate the APC amd ATEPC dataset using aspect extractor provided by PyABSA, see demos/aspect_text_extraction/extract_aspect_and_make_dataset.py"
+  },
+  "1.16.1": {
+    "1": "fix some DatasetItem"
+  },
+  "1.16.0": {
+    "1": "Fix a checkpoint downloading and inflation bug which prevents loading a checkpoint from huggingface spaces",
+    "2": "Fix a important bug which cause unexpected low performance when performing ATEPC inference for Chinese language (and possibly other non-latin languages)",
+    "3": "Minor modifications"
+  },
+  "1.15.7": {
+    "1": "Update findfile dependency to 1.7.9.5",
+    "2": "Default to activate use_bert_spc for ATEPC models to improve ATE and APC performance",
+    "3": "Minor fixes"
+  },
+  "1.15.6": {
+    "1": "Add classification report (including precision, recall, F1) display after training, set config.show_metric to activate",
+    "2": "Add offline option to load huggingface model in inference: set get_xxx(offline=True) to auto detect and load local pretrained model",
+    "3": "Fix lcf-fusion in fast-lsa-t-v2",
+    "4": "Fix some typos",
+    "5": "Rename MOOC-En dataset to MOOC_En"
+  },
+  "1.15.5": {
+    "1": "Minor fixes of tad inference"
+  },
+  "1.15.4": {
+    "1": "Fix some bugs",
+    "2": "Add some new features",
+    "3": "Add a english MOOC dataset"
+  },
+  "1.15.0": {
+    "1": "Fix some bugs"
+  },
+  "1.14.8": {
+    "1": "Minor fixes"
+  },
+  "1.14.7": {
+    "1": "Refactor optimizer support, you can use torch optimizer either by a string or an optimizer object from torch.optim",
+    "2": "Fix checkpoint saving in some scenarios",
+    "3": "Minor fixes"
+  },
+  "1.14.6": {
+    "1": "Minor fixes"
+  },
+  "1.14.5": {
+    "1": "Revert Python version dependency",
+    "2": "Revert Torch dependency"
+  },
+  "1.14.4": {
+    "1": "This version contain breaking experimental changes, if you find any bug please roll back and report on Github"
+  },
+  "1.14.3(2)": {
+    "1": "Refactor pre-tokenization before inference for multilingual ATEPC",
+    "2": "Reset default checkpoint host to Huggingface Hub (useful for Chinese users), please test and report if it works",
+    "3": "Other bug fixes and improvement, see source code"
+  },
+  "1.14.1": {
+    "1": "Minor update"
+  },
+  "1.14.0": {
+    "1": "Bug fixes"
+  },
+  "1.13.4": {
+    "1": "Refactor output save format",
+    "2": "Register SemEval2016Task5 datasets in PyABSA",
+    "3": "Add More language support for ATEPC"
+  },
+  "1.13.3": {
+    "1": "Improve quality of aspect term extraction results"
+  },
+  "1.13.0(1,2)": {
+    "1": "Bug Fixes"
+  },
+  "1.10.6": {
+    "1": "Fix latent resource warning"
+  },
+  "1.10.5": {
+    "1": "Fix config check function",
+    "2": "Fix inference of baseline APC models",
+    "3": "Some refactor, run tests and improve stability",
+    "4": "Remove SSW APC models"
+  },
   "1.10.4": {
     "1": "Fix a bug in dataset detection, which may cause unexpected dataset mis-detection"
   },
@@ -458,4 +558,4 @@
     "2": "Add polarity label-fix features for some datasets containing negative labels",
     "3": "Some typo-fix"
   }
-}
+}