Skip to content

Commit

Permalink
Merge branch 'main' into reorder-callbacks
Browse files Browse the repository at this point in the history
  • Loading branch information
irenedea authored Sep 9, 2024
2 parents ab493fc + e8eca4f commit 2ee540f
Show file tree
Hide file tree
Showing 11 changed files with 14 additions and 10 deletions.
4 changes: 4 additions & 0 deletions llmfoundry/models/layers/blocks.py
Original file line number Diff line number Diff line change
Expand Up @@ -170,7 +170,9 @@ def forward(
extra_kwargs = {}
if prev_layer_key_value is not None:
extra_kwargs['prev_layer_key_value'] = prev_layer_key_value
if key_value_states is not None:
extra_kwargs['key_value_states'] = key_value_states

if self.fuse_norm_attn_norm:
x, m, attn_weights, past_key_value = self.norm_attn_norm(
x,
Expand Down Expand Up @@ -336,7 +338,9 @@ def forward(
extra_kwargs = {}
if prev_layer_key_value is not None:
extra_kwargs['prev_layer_key_value'] = prev_layer_key_value
if key_value_states is not None:
extra_kwargs['key_value_states'] = key_value_states

b, attn_weights, past_key_value = self.attn(
a,
past_key_value=past_key_value,
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-1b-eval.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ integrations:
command: |
cd llm-foundry/scripts/
composer eval/eval.py /mnt/config/parameters.yaml
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: mpt-1b-eval

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-1b-max-seq-len-8k.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ command: |
--out_root ./my-copy-c4 --splits train_small val_small \
--concat_tokens 8192 --tokenizer EleutherAI/gpt-neox-20b --eos_text '<|endoftext|>'
composer train/train.py /mnt/config/parameters.yaml
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: mpt-1b-ctx-8k-gpus-8

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-1b.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ command: |
eval_loader.dataset.split=val_small \
max_duration=100ba \
eval_interval=0
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: mpt-1b-gpus-8

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-benchmark-mpt.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ compute:
# cluster: TODO # Name of the cluster to use for this run
# gpu_type: a100_80gb # Type of GPU to use. We use a100_80gb in our experiments

image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest

integrations:
- integration_type: git_repo
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-convert-composer-to-hf.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@ command: |
--hf_output_path s3://bucket/folder/hf/ \
--output_precision bf16 \
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: convert-composer-hf

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-hf-eval.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ gpu_num: 8
# gpu_type:
# cluster: # replace with your cluster here!

image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest

# The below is injected as a YAML file: /mnt/config/parameters.yaml
parameters:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-hf-generate.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@ command: |
"Here's a quick recipe for baking chocolate chip cookies: Start by" \
"The best 5 cities to visit in Europe are"
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: hf-generate

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-llama2-finetune.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@ integrations:
command: |
cd llm-foundry/scripts
composer train/train.py /mnt/config/parameters.yaml
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
name: llama2-finetune

compute:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-openai-eval.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ gpu_num: #
gpu_type: #
cluster: # replace with your cluster here!

image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest

# The below is injected as a YAML file: /mnt/config/parameters.yaml
parameters:
Expand Down
2 changes: 1 addition & 1 deletion mcli/mcli-pretokenize-oci-upload.yaml
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
name: c4-2k-pre-tokenized
image: mosaicml/llm-foundry:2.4.0_cu124-latest
image: mosaicml/llm-foundry:2.3.1_cu121-latest
compute:
gpus: 8 # Number of GPUs to use

Expand Down

0 comments on commit 2ee540f

Please sign in to comment.