Skip to content

BERT Support #56

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

BERT Support #56

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6fa708e
add muon and sgd_momentum optimizers
Quentin-Anthony Jul 14, 2025
61c45b9
add bert and clear up flag
Quentin-Anthony Jul 29, 2025
d8a3573
add bert and clear up flag
Quentin-Anthony Jul 29, 2025
173aacb
Merge branch 'qanthony/bert' of https://github.com/EleutherAI/cookboo…
Quentin-Anthony Jul 29, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 12 additions & 3 deletions calc/calc_transformer_flops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,7 +65,7 @@ def config_parser():
type=float,
default=300e9,
help='Number of tokens you are training over')
parser.add_argument("--no-checkpoint-activations", "-ca",
parser.add_argument("--no-checkpoint-activations", "-nca",
action='store_false',
help='Whether Megatron-style activation checkpointing is being used',
dest='checkpoint_activations')
Expand All @@ -85,6 +85,13 @@ def config_parser():
parser.add_argument("--infer", "-i",
action='store_true',
help='Pass to calculate FLOPs for inference-only workload (no backward pass)')
parser.add_argument("--encoder-only", "-enc", action="store_true",
help="Set if the model is encoder-only (e.g. BERT/ModernBERT)")

parser.add_argument("--mlm-ratio", "-mr", type=float, default=1.0,
help="Fraction of tokens that receive a language-model head. "
"Use 1.0 for autoregressive GPT-style training, "
"0.15 for BERT-style masked-LM pre-training")
return parser

# calculates the flops of a model given its hparams
Expand All @@ -93,6 +100,8 @@ def calc_flops(args):
assert args.num_layers % args.expert_interval == 0, "Require for simplicity that we don't have hanging dense layers"
assert not args.ffn_hidden_size or (args.ffn_expansion_factor == 4), "both '--ffn-hidden-size' and non-default '-ff' values were specified, these cannot conflict"

is_encoder = args.encoder_only

# An A_(m x k) X B_(k x n) matrix multiplication requires 2m x k x n FLOPs (factor of 2 needed to account for multiplies and adds)

# determine the flops factor.
Expand All @@ -117,8 +126,8 @@ def calc_flops(args):
ffn_flops = int(iter_factor * 2 * args.num_mlp_linears * args.ffn_expansion_factor) * args.num_layers * args.tokens * args.hidden_size * args.hidden_size

# no activation checkpointing for embeddings
# embedding (4*d_model) plus unembedding (2*d_model*vocab_size)
embedding_flops = args.tokens * (4 * args.hidden_size + 2 * args.hidden_size * args.vocab_size)
embedding_flops = (args.tokens * (4 * args.hidden_size + 2 * args.hidden_size * args.vocab_size)
* (args.mlm_ratio if is_encoder else 1.0))

if args.moe and args.topk > 1:
ffn_flops *= args.topk / args.expert_interval
Expand Down