vwxyzjn

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
import torch.nn.functional as F

model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM2-135M")
tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/SmolLM2-135M")
tokenizer.add_special_tokens({"pad_token": "<PAD>"})
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
device = torch.device("cpu")
model.to(device)

vwxyzjn

import argparse
import numpy as np


p = 100 # padding token id
o = 1 # observation (prompt / input ids)
a = 2 # action (response ids)

queries = [
    [p, p, o, o, o],

vwxyzjn

import torch
import torch.nn as nn
import torch.optim as optim

# Create target distribution (fixed)
target_logits = torch.randn(10)
target_log_probs = torch.log_softmax(target_logits, dim=0)

# Create learnable distribution
learnable_logits = nn.Parameter(torch.rand_like(target_logits))  # Initialize randomly

vwxyzjn

{
  "name": "material-ui-nextjs-ts",
  "version": "5.0.0",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "material-ui-nextjs-ts",
      "version": "5.0.0",
      "dependencies": {

vwxyzjn

# Taken and modified from https://github.com/huggingface/trl
# Copyright 2024 The AllenAI Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software

vwxyzjn

import json
import os
os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "0"

import gradio as gr
import numpy as np
import pandas as pd
from apscheduler.schedulers.background import BackgroundScheduler
from huggingface_hub import HfApi, snapshot_download

vwxyzjn

# coding=utf-8
# Adapted from
# https://github.com/huggingface/transformers/blob/v4.40.1/src/transformers/models/olmo/modeling_olmo.py
# Copyright 2024 The vLLM team.
# Copyright 2024 EleutherAI and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.

vwxyzjn

python scripts/submit_finetune_job.py \
    --cluster ai2/augusta-google-1 \
    --priority high \
    --workspace ai2/tulu-3-dev \
    --num_nodes 4 \
    --image costah/open_instruct_ppo_ray_ninja \
    --default_beaker_config configs/beaker_configs/default_finetune_multinode.yaml \
    --config configs/train_configs/sft/tulu3_8b_preview_mix_v3.9.yaml \
    --exp_name olmo1124_finetune

vwxyzjn

# coding=utf-8
# Adapted from
# https://github.com/huggingface/transformers/blob/v4.40.1/src/transformers/models/olmo/modeling_olmo.py
# Copyright 2024 The vLLM team.
# Copyright 2024 EleutherAI and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.

vwxyzjn

# coding=utf-8
# Adapted from
# https://github.com/huggingface/transformers/blob/v4.40.1/src/transformers/models/olmo/modeling_olmo.py
# Copyright 2024 The vLLM team.
# Copyright 2024 EleutherAI and the HuggingFace Inc. team. All rights reserved.
#
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
# and OPT implementations in this library. It has been modified from its
# original forms to accommodate minor architectural differences compared
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.