metadata
dataset_info:
features:
- name: weight_type
dtype: string
- name: P_w
list: float64
- name: sum
dtype: float32
- name: mean
dtype: float32
- name: std
dtype: float32
- name: max
dtype: float32
- name: min
dtype: float32
- name: skew
dtype: float32
- name: kurtosis
dtype: float32
- name: differential_entropy
dtype: float32
- name: entropy
dtype: float64
- name: fit_mu
dtype: float64
- name: fit_sigma
dtype: float64
- name: kl_vs_empirical_normal
dtype: float64
- name: SVD
list: float32
- name: P_sv
list: float64
- name: head
dtype: int64
- name: layer
dtype: int64
- name: model
dtype: string
- name: revision
dtype: string
- name: step
dtype: int64
- name: job_uuid
dtype: string
- name: job_id
dtype: string
splits:
- name: train
num_bytes: 4799596040
num_examples: 473088
download_size: 212650723
dataset_size: 4799596040
configs:
- config_name: default
data_files:
- split: train
path: data/train-*