metadata

license: apache-2.0
tags:
  - summarization
  - t5
datasets:
  - billsum
metrics:
  - rouge
widget:
  - text: >-
      The people of the State of California do enact as follows: SECTION 1. The
      Legislature hereby finds and declares as follows: (a) Many areas of the
      state are disproportionately impacted by drought because they are heavily
      dependent or completely reliant on groundwater from basins that are in
      overdraft and in which the water table declines year after year or from
      basins that are contaminated. (b) There are a number of state grant and
      loan programs that provide financial assistance to communities to address
      drinking water and wastewater needs. Unfortunately, there is no program in
      place to provide similar assistance to individual homeowners who are
      reliant on their own groundwater wells and who may not be able to afford
      conventional private loans to undertake vital water supply, water quality,
      and wastewater improvements. (c) The program created by this act is
      intended to bridge that gap by providing low-interest loans, grants, or
      both, to individual homeowners to undertake actions necessary to provide
      safer, cleaner, and more reliable drinking water and wastewater treatment.
      These actions may include, but are not limited to, digging deeper wells,
      improving existing wells and related equipment, addressing drinking water
      contaminants in the homeowner’s water, or connecting to a local water or
      wastewater system. SEC. 2. Chapter 6.6 (commencing with Section 13486) is
      added to Division 7 of the Water Code, to read: CHAPTER  6.6. Water and
      Wastewater Loan and Grant Program 13486. (a) The board shall establish a
      program in accordance with this chapter to provide low-interest loans and
      grants to local agencies for low-interest loans and grants to eligible
      applicants for any of the following purposes:
    example_title: Water use
  - text: >-
      The people of the State of California do enact as follows:   SECTION 1.
      Section 2196 of the Elections Code is amended to read: 2196. (a) (1)
      Notwithstanding any other provision of law, a person who is qualified to
      register to vote and who has a valid California driver’s license or state
      identification card may submit an affidavit of voter registration
      electronically on the Internet Web site of the Secretary of State. (2) An
      affidavit submitted pursuant to this section is effective upon receipt of
      the affidavit by the Secretary of State if the affidavit is received on or
      before the last day to register for an election to be held in the precinct
      of the person submitting the affidavit. (3) The affiant shall
      affirmatively attest to the truth of the information provided in the
      affidavit. (4) For voter registration purposes, the applicant shall
      affirmatively assent to the use of his or her signature from his or her
      driver’s license or state identification card. (5) For each electronic
      affidavit, the Secretary of State shall obtain an electronic copy of the
      applicant’s signature from his or her driver’s license or state
      identification card directly from the Department of Motor Vehicles. (6)
      The Secretary of State shall require a person who submits an affidavit
      pursuant to this section to submit all of the following: (A) The number
      from his or her California driver’s license or state identification card.
      (B) His or her date of birth. (C) The last four digits of his or her
      social security number. (D) Any other information the Secretary of State
      deems necessary to establish the identity of the affiant. (7) Upon
      submission of an affidavit pursuant to this section, the electronic voter
      registration system shall provide for immediate verification of both of
      the following:
    example_title: Election
model-index:
  - name: t5-small-finetuned-billsum-ca_test
    results:
      - task:
          type: text2text-generation
          name: Sequence-to-sequence Language Modeling
        dataset:
          name: billsum
          type: billsum
          args: default
        metrics:
          - type: rouge
            value: 12.6315
            name: Rouge1
      - task:
          type: summarization
          name: Summarization
        dataset:
          name: billsum
          type: billsum
          config: default
          split: test
        metrics:
          - type: rouge
            value: 12.1368
            name: ROUGE-1
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNGYyZjQ3ZDIzNWI5YTcyNmZjNWYwYWNkMzQ1NjVmNmJmMjI3MjIzNmJjODY2ODY5OWE2ODlhZDhlY2QxNmE3OSIsInZlcnNpb24iOjF9.OgYBwYP2BsMkga2iQCaEhwS577_orJZijpxjsd8i-QPOOAes4WiiAmxr5mHrsG7fjp2qXChzbx7Kik-7zpyeBA
          - type: rouge
            value: 4.6017
            name: ROUGE-2
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZWVkN2ZmNGQ1N2QyYzViMzE3ZTc4OWRjMjY5NWQ5OGJmZDg1YTc2NmQyYTYwZTE5MjI0YmEwMGEzMjczZGI1YSIsInZlcnNpb24iOjF9.CGrm9rdFWLMKBE6ZliCgoAIwFlRQ9RrkTTyEXWN-mgs3otZt3PMTHflliDMB2VOzaYqn0AWcb4R6KZSPjveLDg
          - type: rouge
            value: 10.0767
            name: ROUGE-L
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZTkzZGM0NjY4ZTEzNTExZjg3MTU3ODE2MmU3MTc2NjNkZmFhZWVkYzExZTAxOTdhMTczYjU4NzUwNmRlM2FmNCIsInZlcnNpb24iOjF9.6VT0EGxHvnO3-XHbSZ4FtZjRo3IubQm4QFa3Jxt-Wc-avAUqA_emMEhe6CkJHJLaqbalN-pzRL3wVoWO1YUXCw
          - type: rouge
            value: 10.6892
            name: ROUGE-LSUM
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYTM3ZDE2NzE0NWE1ODdlNGEwNGY4ZWZiOTc3OGEwZTU2MmEzODUwMjBjODE1MGZiMTJmOWY2MTYwZDIzZTQ5YSIsInZlcnNpb24iOjF9.DuebV54ulf7s9lnV_0MbbX_8rfXD5eIuwLEYUrqyqdwJwoLdljPtmr4eeJyBZnRWQuYST3X9MP-P7cxeOLVdAA
          - type: loss
            value: 2.897707462310791
            name: loss
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMmU3NmZjZmY3MTI3NTU1MzZiZmQxNGM5YTY2NzE5MDhhMjgyZTZjYmNiN2Q0NGQ3ZTRkYmUwOGE5YWMwOWVmNiIsInZlcnNpb24iOjF9.JOiS9YvTSoP4Ig-QUzZP-VCF0YjUMK-yLTDRuhml_iWnavkvcUFvOYrZR70nV7lufXyGFBXseUDjeFRhYd0GDw
          - type: gen_len
            value: 19
            name: gen_len
            verified: true
            verifyToken: >-
              eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNGFlNzdlMGI1NDgwYTdhYjEyZjcyMWM5MmI3MmQ5MTExZmM3YTU0YTBjMTlkYTNkMjJmMDljNjBmYzAwNzYxYyIsInZlcnNpb24iOjF9.03VYE7XRLy0dRRAD8lljnmSu7mVen-RwglqZ2fpgm8hAkoIXoSeyLbJl3sKcf28lQl1I-40ySEIwoGokfF7ABg

t5-small-finetuned-billsum-ca_test

This model is a fine-tuned version of t5-small on the billsum dataset. It achieves the following results on the evaluation set:

Loss: 2.3376
Rouge1: 12.6315
Rouge2: 6.9839
Rougel: 10.9983
Rougelsum: 11.9383
Gen Len: 19.0

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 4
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	495	2.4805	9.9389	4.1239	8.3979	9.1599	19.0
3.1564	2.0	990	2.3833	12.1026	6.5196	10.5123	11.4527	19.0
2.66	3.0	1485	2.3496	12.5389	6.8686	10.8798	11.8636	19.0
2.5671	4.0	1980	2.3376	12.6315	6.9839	10.9983	11.9383	19.0

Framework versions

Transformers 4.12.2
Pytorch 1.9.0+cu111
Datasets 1.14.0
Tokenizers 0.10.3