Save model at global step 105
Browse files- config.yaml +5 -5
- extra_state_world_size_4_rank_0.pt +1 -1
- extra_state_world_size_4_rank_1.pt +1 -1
- extra_state_world_size_4_rank_2.pt +1 -1
- extra_state_world_size_4_rank_3.pt +1 -1
- model_world_size_4_rank_0.pt +1 -1
- model_world_size_4_rank_1.pt +1 -1
- model_world_size_4_rank_2.pt +1 -1
- model_world_size_4_rank_3.pt +1 -1
- optim_world_size_4_rank_0.pt +1 -1
- optim_world_size_4_rank_1.pt +1 -1
- optim_world_size_4_rank_2.pt +1 -1
- optim_world_size_4_rank_3.pt +1 -1
config.yaml
CHANGED
@@ -61,7 +61,7 @@ actor_rollout_ref:
|
|
61 |
lr_warmup_steps_ratio: 0.0
|
62 |
min_lr_ratio: null
|
63 |
warmup_style: constant
|
64 |
-
total_training_steps:
|
65 |
weight_decay: 0.01
|
66 |
fsdp_config:
|
67 |
wrap_policy:
|
@@ -122,7 +122,7 @@ critic:
|
|
122 |
lr_warmup_steps_ratio: 0.0
|
123 |
min_lr_ratio: null
|
124 |
warmup_style: constant
|
125 |
-
total_training_steps:
|
126 |
weight_decay: 0.01
|
127 |
model:
|
128 |
path: meta-llama/Llama-3.2-3B-Instruct
|
@@ -193,7 +193,7 @@ algorithm:
|
|
193 |
trainer:
|
194 |
balance_batch: true
|
195 |
total_epochs: 100
|
196 |
-
total_training_steps:
|
197 |
project_name: value-LLM
|
198 |
experiment_name: ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256
|
199 |
logger:
|
@@ -218,7 +218,7 @@ trainer:
|
|
218 |
resume_from_hf:
|
219 |
enable: true
|
220 |
actor_hf_repo_id: RyanYr/ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256_actor
|
221 |
-
actor_revision:
|
222 |
critic_hf_repo_id: RyanYr/ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256_critic
|
223 |
-
critic_revision:
|
224 |
hf_token: null
|
|
|
61 |
lr_warmup_steps_ratio: 0.0
|
62 |
min_lr_ratio: null
|
63 |
warmup_style: constant
|
64 |
+
total_training_steps: 200
|
65 |
weight_decay: 0.01
|
66 |
fsdp_config:
|
67 |
wrap_policy:
|
|
|
122 |
lr_warmup_steps_ratio: 0.0
|
123 |
min_lr_ratio: null
|
124 |
warmup_style: constant
|
125 |
+
total_training_steps: 200
|
126 |
weight_decay: 0.01
|
127 |
model:
|
128 |
path: meta-llama/Llama-3.2-3B-Instruct
|
|
|
193 |
trainer:
|
194 |
balance_batch: true
|
195 |
total_epochs: 100
|
196 |
+
total_training_steps: 200
|
197 |
project_name: value-LLM
|
198 |
experiment_name: ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256
|
199 |
logger:
|
|
|
218 |
resume_from_hf:
|
219 |
enable: true
|
220 |
actor_hf_repo_id: RyanYr/ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256_actor
|
221 |
+
actor_revision: ea0afbb1bdbd52f362714b38917a2a8aa44b1193
|
222 |
critic_hf_repo_id: RyanYr/ppo-dapo-llama3.2-3B-it-lr-bs1024-mbs256_critic
|
223 |
+
critic_revision: 431de5f1de160e60af68af4b0829ffe1f041be21
|
224 |
hf_token: null
|
extra_state_world_size_4_rank_0.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14632
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b997bc838bd4ab9241c7f47aad59c5e3557197232c3861495edbd2e063f12af8
|
3 |
size 14632
|
extra_state_world_size_4_rank_1.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14632
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a2339072c15749307da3ff25dd8c1a13b5e202c75d6786d32bd5c855efc7b8a5
|
3 |
size 14632
|
extra_state_world_size_4_rank_2.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14632
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1f0d0f28ab9e9f6049827a3a88ab60becb6659addef59a8ba4f71f0be0f195a1
|
3 |
size 14632
|
extra_state_world_size_4_rank_3.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14632
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:133bf50a541947ff52c35ddcc3899c3343d303719b8e2e8339f5913cb51e7b8d
|
3 |
size 14632
|
model_world_size_4_rank_0.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3212915290
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0570adab62350fa4271a71052e61a8491f8c1a61ccbb2aefd89bac4156097127
|
3 |
size 3212915290
|
model_world_size_4_rank_1.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3212903002
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4f1c99f08566c14bb7566106d4fffac48d9189418ac82846538737b3a6866517
|
3 |
size 3212903002
|
model_world_size_4_rank_2.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3212903002
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:764aec12567c84c4ace83f46a8a0f465b5e231a35919490fa562282cff91cbe5
|
3 |
size 3212903002
|
model_world_size_4_rank_3.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3212903002
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1b20f8ff3f2e37ad8133b3a54cfb1091b56d192a4c92284dba0b762333f0421a
|
3 |
size 3212903002
|
optim_world_size_4_rank_0.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6425532375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:23b2d3d57460afe6d47897fda6f1f718a6844ab01e7b08dc2b14adbd0c874977
|
3 |
size 6425532375
|
optim_world_size_4_rank_1.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6425532375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0bb91d67c3653fc876857f9d0982da447d06f4e2d0afbaf11e1515ae0b4289c0
|
3 |
size 6425532375
|
optim_world_size_4_rank_2.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6425532375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:20c3780af420df59396622c84c35242f9f800451a6484aeb49b48ceda1ad355b
|
3 |
size 6425532375
|
optim_world_size_4_rank_3.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6425532375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ae73c2c8b387c38e7d16704e25b0025f07d77398668327b29e97b1dfceb39bb2
|
3 |
size 6425532375
|