Update README.md
Browse files
README.md
CHANGED
@@ -20,7 +20,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
20 |
|
21 |
# output
|
22 |
|
23 |
-
This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on an LegalBench dataset.
|
24 |
It achieves the following results on the evaluation set:
|
25 |
- Loss: 1.1885
|
26 |
- Rouge1: 65.4762
|
@@ -31,7 +31,17 @@ It achieves the following results on the evaluation set:
|
|
31 |
|
32 |
## Model description
|
33 |
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
|
36 |
## Intended uses & limitations
|
37 |
|
@@ -39,10 +49,12 @@ More information needed
|
|
39 |
|
40 |
## Training and evaluation data
|
41 |
|
42 |
-
|
43 |
|
44 |
## Training procedure
|
45 |
|
|
|
|
|
46 |
### Training hyperparameters
|
47 |
|
48 |
The following hyperparameters were used during training:
|
|
|
20 |
|
21 |
# output
|
22 |
|
23 |
+
This model is a fine-tuned version of [google/flan-t5-base](https://huggingface.co/google/flan-t5-base) on an [LegalBench](https://github.com/HazyResearch/legalbench) dataset.
|
24 |
It achieves the following results on the evaluation set:
|
25 |
- Loss: 1.1885
|
26 |
- Rouge1: 65.4762
|
|
|
31 |
|
32 |
## Model description
|
33 |
|
34 |
+
We finetune [Flan-T5-Base]((https://huggingface.co/google/flan-t5-base)) LLM on the [LegalBench](https://github.com/HazyResearch/legalbench).
|
35 |
+
|
36 |
+
### Prompt
|
37 |
+
The prompt should be formatted as follows:
|
38 |
+
{{Question}} {{Clause}}
|
39 |
+
|
40 |
+
Question: Does the clause grant one party an “enterprise,” “all you can eat” or unlimited usage license?
|
41 |
+
|
42 |
+
Clause: Except as the parties may otherwise agree in writing, Converge, to the extent it has the legal right to do so, hereby grants to Vert an irrevocable, perpetual, world-wide, non-exclusive right and license to use, load, store, transmit, execute, copy, market, distribute, in any medium or distribution technology whatsoever, known or unknown, display, perform and sublicense the Converge-Independent Materials and the Third-Party Materials, in both Source Code and Object Code formats, and to make unlimited instantiations thereof, for any and all purposes.
|
43 |
+
|
44 |
+
Prompt: Does the clause grant one party an “enterprise,” “all you can eat” or unlimited usage license? Except as the parties may otherwise agree in writing, Converge, to the extent it has the legal right to do so, hereby grants to Vert an irrevocable, perpetual, world-wide, non-exclusive right and license to use, load, store, transmit, execute, copy, market, distribute, in any medium or distribution technology whatsoever, known or unknown, display, perform and sublicense the Converge-Independent Materials and the Third-Party Materials, in both Source Code and Object Code formats, and to make unlimited instantiations thereof, for any and all purposes.
|
45 |
|
46 |
## Intended uses & limitations
|
47 |
|
|
|
49 |
|
50 |
## Training and evaluation data
|
51 |
|
52 |
+
We used [LegalBench](https://github.com/HazyResearch/legalbench) for training and evaluation.
|
53 |
|
54 |
## Training procedure
|
55 |
|
56 |
+
Tutorial: Finetune [Flan-T5](https://docs.blueprint.baseten.co/finetuning/flan-t5/) with Baseten.
|
57 |
+
|
58 |
### Training hyperparameters
|
59 |
|
60 |
The following hyperparameters were used during training:
|