Enable multiple LoRa adapters #2010
drbh
force pushed
from
091f2dce
to
d103264e
1 year ago
drbh
marked this pull request as ready for review 1 year ago
drbh
changed the title Lora internal Enable multiple LoRa adapters 1 year ago
feat: first draft load multiple lora
db3d8e65
feat: load weights within layer and refactor lora pass
0a6ea7fb
fix: refactor and reduce lora math
a046c303
feat: baseline impl single request multi lora support
c6616312
feat: prefer lorax implementation and port loading logic
8b50f4b7
fix: prefer adapter_data and refactors
d5f21d57
feat: perfer loraxs custom punica kernels and add mlp loras
8984ce6c
fix: adjust batch for bgmv
ad088d51
fix: adjust adapter_segments logic when in batch
c9273767
fix: refactor and move changes to v3 proto
73eb2ae2
fix: pass model_id for all flash causal lms
88bd5c2c
fix: pass model_id for all causal and seq2seq lms
dc0f7655
fix: add model_id to model test
9c45d349
feat: add lora support to mistral and refactors
de56a81c
feat: prefer model id in request
68399c1a
fix: include rust code for adapter id
81707bfb
feat: bump launcher and add new lora docs
43ec9dfe
feat: support base model generation and refactors
611225f0
fix: rename doc to retry ci build
a563a931
feat: support if vlm models
91f40722
fix: add adapter_data param and avoid missing layers
b1169273
fix: add adapter_data param to phi and neox
1deb3725
fix: update all models forwards to include adapter_data
101b95ad
fix: add model_id to IdeficsCausalLM
ce40ad26
Update lora.md
1be1ebc4
Update lora.md
d6cf63ca
drbh
force pushed
from
5a0ed2b3
to
d6cf63ca
1 year ago
fix: add lora kernel to dockerfile, support running without kernels a…
aa88c4fd
fix: avoid dockerfile conflict
06c3254c
fix: merge 'main' into lora-internal to resolve conflicts
0e1c28ca
Merge branch 'main' into lora-internal
1104885f
Merge branch 'main' into lora-internal
224455f3
fix: refactors and adjust flash llama lora logic
4f1543d3
fix: skip llama test due to CI issue (temp)
ce70fce9
fix: skip llama test CI (temp) 2
c9e4526b
fix: revert skips and prefer updated ci token for tests
a07b6129
fix: refactors and helpful comments
3c9b28ea
fix: add noop in TensorParallelAdapterRowLinear too
c927cffb
fix: refactor and move shard_lora_weights logic
f94f2b3e
Merge branch 'main' into lora-internal
0d496baa
danieldk
dismissed these changes
on 2024-06-25
fix: exit early if no adapter_data
a2d821c4
drbh
dismissed their stale review
via a2d821c4
1 year ago
Merge branch 'main' into lora-internal
59575fe6
drbh
merged
04e1af94
into main 1 year ago
drbh
deleted the lora-internal branch 1 year ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub