q transformer Download - q transformer Quellcode herunterladen

q transformer

AI-Quellcode

0.3.0

Herunterladen

Q-Transformator

Implementierung von Q-Transformer, skalierbarem Offline-Reinforcement-Lernen über autoregressive Q-Funktionen, aus Google Deepmind

Ich werde die Logik für Q-Learning für einzelne Aktionen beibehalten, nur um einen abschließenden Vergleich mit dem vorgeschlagenen autoregressiven Q-Learning für mehrere Aktionen zu ermöglichen. Auch um der Aufklärung für mich und die Öffentlichkeit zu dienen.

Die autoregressive Q-Learning-Formulierung wurde von Kotb et al. reproduziert.

Installieren

$ pip install q-transformer

Verwendung

 import torch

from q_transformer import (
    QRoboticTransformer ,
    QLearner ,
    Agent ,
    ReplayMemoryDataset
)

# the attention model

model = QRoboticTransformer (
    vit = dict (
        num_classes = 1000 ,
        dim_conv_stem = 64 ,
        dim = 64 ,
        dim_head = 64 ,
        depth = ( 2 , 2 , 5 , 2 ),
        window_size = 7 ,
        mbconv_expansion_rate = 4 ,
        mbconv_shrinkage_rate = 0.25 ,
        dropout = 0.1
    ),
    num_actions = 8 ,
    action_bins = 256 ,
    depth = 1 ,
    heads = 8 ,
    dim_head = 64 ,
    cond_drop_prob = 0.2 ,
    dueling = True
)

# you need to supply your own environment, by overriding BaseEnvironment

from q_transformer . mocks import MockEnvironment

env = MockEnvironment (
    state_shape = ( 3 , 6 , 224 , 224 ),
    text_embed_shape = ( 768 ,)
)

# env.init()     should return instructions and initial state: Tuple[str, Tensor[*state_shape]]
# env(actions)   should return rewards, next state, and done flag: Tuple[Tensor[()], Tensor[*state_shape], Tensor[()]]

# agent is a class that allows the q-model to interact with the environment to generate a replay memory dataset for learning

agent = Agent (
    model ,
    environment = env ,
    num_episodes = 1000 ,
    max_num_steps_per_episode = 100 ,
)

agent ()

# Q learning on the replay memory dataset on the model

q_learner = QLearner (
    model ,
    dataset = ReplayMemoryDataset (),
    num_train_steps = 10000 ,
    learning_rate = 3e-4 ,
    batch_size = 4 ,
    grad_accum_every = 16 ,
)

q_learner ()

# after much learning
# your robot should be better at selecting optimal actions

video = torch . randn ( 2 , 3 , 6 , 224 , 224 )

instructions = [
    'bring me that apple sitting on the table' ,
    'please pass the butter'
]

actions = model . get_optimal_actions ( video , instructions )

Anerkennung

StabilityAI, A16Z Open Source AI Grant Program und ? Huggingface für die großzügigen Sponsoren und auch für meine anderen Sponsoren, die mir die Unabhängigkeit gegeben haben, aktuelle Forschung im Bereich der künstlichen Intelligenz als Open-Source-Quelle zu nutzen

Todo

Zitate

 @inproceedings { qtransformer ,
    title   = { Q-Transformer: Scalable Offline Reinforcement Learning via Autoregressive Q-Functions } ,
    authors = { Yevgen Chebotar and Quan Vuong and Alex Irpan and Karol Hausman and Fei Xia and Yao Lu and Aviral Kumar and Tianhe Yu and Alexander Herzog and Karl Pertsch and Keerthana Gopalakrishnan and Julian Ibarz and Ofir Nachum and Sumedh Sontakke and Grecia Salazar and Huong T Tran and Jodilyn Peralta and Clayton Tan and Deeksha Manjunath and Jaspiar Singht and Brianna Zitkovich and Tomas Jackson and Kanishka Rao and Chelsea Finn and Sergey Levine } ,
    booktitle = { 7th Annual Conference on Robot Learning } ,
    year   = { 2023 }
}

 @inproceedings { dao2022flashattention ,
    title   = { Flash{A}ttention: Fast and Memory-Efficient Exact Attention with {IO}-Awareness } ,
    author  = { Dao, Tri and Fu, Daniel Y. and Ermon, Stefano and Rudra, Atri and R{'e}, Christopher } ,
    booktitle = { Advances in Neural Information Processing Systems } ,
    year    = { 2022 }
}

 @inproceedings { Kumar2023MaintainingPI ,
    title   = { Maintaining Plasticity in Continual Learning via Regenerative Regularization } ,
    author  = { Saurabh Kumar and Henrik Marklund and Benjamin Van Roy } ,
    year    = { 2023 } ,
    url     = { https://api.semanticscholar.org/CorpusID:261076021 }
}