update docs for parse_t5 v2

README.md

@@ -34,7 +34,7 @@ The system includes different neural-network models for parsing and for generati
 [SPRING model](https://github.com/SapienzaNLP/spring). The basic process is similar to the parse_t5 model except that
 the pretrained transformer used is the bart-large model which is larger than the t5-base model used in parse_t5.
 
-* Parse (StoG) model parse_t5 gives an **81 SMATCH score** with LDC2020T02.  This model uses the
+* Parse (StoG) model parse_t5 gives an **82 SMATCH score** with LDC2020T02.  This model uses the
 pretrained HuggingFace T5 transformer model to convert sentences to graph-encoded sequences which
 are then deserialized into an AMR graph.
 

docs/install.md

@@ -43,7 +43,7 @@ Download the pretrained parse models from:
 * [model_parse_spring-v0_1_0.tar.gz](https://github.com/bjascob/amrlib-models/releases/download/model_parse_spring-v0_1_0/model_parse_spring-v0_1_0.tar.gz)
 (or see [GitHub/amrlib-models](https://github.com/bjascob/amrlib-models))
 
-* [model_parse_t5-v0_1_0.tar.gz](https://github.com/bjascob/amrlib-models/releases/download/model_parse_t5-v0_1_0/model_parse_t5-v0_1_0.tar.gz)
+* [model_parse_t5-v0_2_0.tar.gz](https://github.com/bjascob/amrlib-models/releases/download/model_parse_t5-v0_2_0/model_parse_t5-v0_2_0.tar.gz)
 (or see [GitHub/amrlib-models](https://github.com/bjascob/amrlib-models))
 
 * [model_parse_gsii-v0_1_0.tar.gz](https://u.pcloud.link/publink/show?code=XZD2z0XZOqRtS2mNMHhMG4UhXOCNO4yzeaLk)

docs/models.md

@@ -19,7 +19,7 @@ Additional inference parameters:
 See amrlib/models/parse_spring/inference.py for implementation details.
 
 ## Parse T5
-**81 SMATCH score** with LDC2020T02.
+**82 SMATCH score** with LDC2020T02.
 
 This model is based on the pretrained [HuggingFace](https://github.com/huggingface/transformers)
 T5 transformer to convert sentences to graph-encoded sequences which are then deserialized into