Skip to content
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.

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

bfloat16 support, and an attempt at homogenizing model_dtype & precision #54

Merged
merged 17 commits into from
Jul 10, 2024
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
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
Prev Previous commit
update cometkiwi recipe
  • Loading branch information
francoishernandez committed Jul 10, 2024
commit 7e1a532d2fd402c1bde9c63d213f0839cbf10778
2 changes: 1 addition & 1 deletion recipes/cometkiwi/cometkiwi-xxl-eole.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,7 @@ training:
accum_steps: [0]

# Optimization
model_dtype: "fp16"
compute_dtype: "fp16"
#apex_opt_level: "O2"
optim: "fusedadam"
learning_rate: 0.000015
Expand Down
7 changes: 4 additions & 3 deletions recipes/mistral/mistral-7b-awq-gemm-inference.yaml
Original file line number Diff line number Diff line change
@@ -1,11 +1,11 @@
transforms: [sentencepiece]
transforms_configs:
sentencepiece:
src_subword_model: "$EOLE_MODEL_DIR/mistral-7b-instruct-v0.2-awq/tokenizer.model"
tgt_subword_model: "$EOLE_MODEL_DIR/mistral-7b-instruct-v0.2-awq/tokenizer.model"
src_subword_model: "$EOLE_MODEL_DIR/mistral-7b-v0.3/tokenizer.model"
tgt_subword_model: "$EOLE_MODEL_DIR/mistral-7b-v0.3/tokenizer.model"

# Model info
model_path: "$EOLE_MODEL_DIR/mistral-7b-instruct-v0.2-awq"
model_path: "$EOLE_MODEL_DIR/mistral-7b-v0.3"

# Inference
seed: 42
Expand All @@ -20,6 +20,7 @@ gpu_ranks: [0]
# parallel_mode: "tensor_parallel"
#quant_layers: ['gate_up_proj', 'down_proj', 'up_proj', 'linear_values', 'linear_query', 'linear_keys', 'final_linear']
#quant_type: "bnb_NF4"
# precision: fp16
precision: fp16
#random_sampling_topk: 1
#random_sampling_topp: 0.6
Expand Down
Loading