issues
search
whittle-org
/
whittle
Python library to compress LitGPT models for resource efficient inference.
https://whittle-org.github.io/whittle/latest/
Apache License 2.0
10
stars
4
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Support initializing a smaller dense model from weights of a larger model
#177
rheasukthanker
opened
13 hours ago
0
fix: addition and standardization of docstrings
#176
zeqri
opened
2 days ago
0
support for structural pruning methods
#175
aaronkl
opened
4 days ago
3
ci: add python 3.12 support
#174
snikhil24
closed
4 days ago
2
feat: adds workflow for pretraining a super-network
#173
aaronkl
opened
5 days ago
0
fix: set input variables as required positional arguments
#172
zeqri
closed
5 days ago
0
add workflow for multi-objective search
#171
aaronkl
opened
6 days ago
0
add workflow for structural pruning
#170
aaronkl
closed
22 hours ago
1
add workflow for distillation
#169
aaronkl
opened
6 days ago
0
fix: update tokenizers in pyproject.toml
#168
rheasukthanker
closed
1 week ago
0
fix: fixed type hints
#167
zeqri
closed
1 week ago
0
feat: add python 3.12 support
#166
timurcarstensen
closed
4 days ago
3
Change sub_network_head_size and query_groups to positional arguments
#165
gabikadlecova
closed
5 days ago
1
Update doc strings
#164
rheasukthanker
opened
1 week ago
3
fixed type hints
#163
zeqri
closed
1 week ago
1
feat: allows to pass fabric
#162
aaronkl
closed
1 week ago
0
Adapt code for latest tokenizers version
#161
rheasukthanker
closed
1 week ago
0
fix: deprecate flexible mlp heads
#160
rheasukthanker
closed
1 week ago
0
Support hw-gpt-bench
#159
rheasukthanker
opened
2 weeks ago
0
feat: Flexible intermediate sizes
#158
gabikadlecova
closed
2 weeks ago
0
adapt litgpt workflow for pre-training
#157
aaronkl
opened
2 weeks ago
0
Remove redundant cos, sin assignment in ```max_seq_length```
#156
rheasukthanker
opened
2 weeks ago
0
fix: branch protection rules interfere with release workflow
#155
timurcarstensen
opened
2 weeks ago
0
fix: forcing deepspeed to use CPU for profiling FLOPS
#154
timurcarstensen
closed
1 week ago
4
deepspeed on gpus
#153
rheasukthanker
closed
1 week ago
2
fix: refactor names of metric
#152
aaronkl
closed
3 weeks ago
0
fix: Extract weights for norm layers, test with random initialization.
#151
gabikadlecova
closed
3 weeks ago
0
Fix norms and randomize tests for subnet extraction
#150
rheasukthanker
closed
3 weeks ago
0
ci: switching to `uv` for github actions
#149
timurcarstensen
closed
3 weeks ago
0
use `uv` for github actions workflows
#148
timurcarstensen
closed
3 weeks ago
0
feat: add support for LLamaMLP in extract_sub_network
#147
gabikadlecova
closed
3 weeks ago
1
use fixed number of heads / intermediate size per layer
#146
aaronkl
closed
1 week ago
2
feat: adding flops and macs profiling for subnets
#145
timurcarstensen
closed
3 weeks ago
0
fix: rename call function
#144
aaronkl
closed
3 weeks ago
0
fix: handle device in GPT model properly
#143
aaronkl
closed
3 weeks ago
0
unify naming of metrics
#142
aaronkl
closed
3 weeks ago
0
feat: add script to profile latency
#141
rheasukthanker
closed
4 weeks ago
0
fix: delete supernet_configs directory
#140
rheasukthanker
closed
4 weeks ago
0
Support extract_sub_network for other MLP classes
#139
gabikadlecova
closed
3 weeks ago
4
Add support for only_predict in evaluation function
#138
arberzela
opened
1 month ago
0
Supernets with varying block sizes
#137
gabikadlecova
closed
2 weeks ago
8
Compute FLOPs of sub-networks
#136
aaronkl
closed
3 weeks ago
0
chore: renaming sinc to sine_curves
#135
timurcarstensen
closed
1 month ago
0
Rename `sinc` example to `sine_curves`
#134
timurcarstensen
closed
1 month ago
0
fix: deprecate sample_random_indices
#133
rheasukthanker
closed
1 month ago
0
Remove sample_random_indices
#132
rheasukthanker
closed
1 month ago
0
feat: modify rope for llama-3 and support llama-3.2
#131
rheasukthanker
closed
1 month ago
0
Support Meta-llama-3.2
#130
rheasukthanker
closed
1 month ago
0
Match llama-3.1-8B results (using lm-eval-harness) from litgpt/huggingface api
#129
rheasukthanker
closed
1 month ago
0
Profile subnet latency
#128
rheasukthanker
closed
4 weeks ago
1
Next