romanri85 commited on
Commit
463ede3
·
verified ·
1 Parent(s): fdae3f4

Upload folder using huggingface_hub

Browse files
epoch225/adapter_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": null,
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": false,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": false,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 32,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.0,
22
+ "megatron_config": null,
23
+ "megatron_core": "megatron.core",
24
+ "modules_to_save": null,
25
+ "peft_type": "LORA",
26
+ "peft_version": "0.18.1",
27
+ "qalora_group_size": 16,
28
+ "r": 32,
29
+ "rank_pattern": {},
30
+ "revision": null,
31
+ "target_modules": [
32
+ "k",
33
+ "ffn.0",
34
+ "v",
35
+ "q",
36
+ "o",
37
+ "ffn.2"
38
+ ],
39
+ "target_parameters": null,
40
+ "task_type": null,
41
+ "trainable_token_indices": null,
42
+ "use_dora": false,
43
+ "use_qalora": false,
44
+ "use_rslora": false
45
+ }
epoch225/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c85188cdbc4ff87d456d560960134cff68cd8dc18f36bed9c1400ba4fcd3537
3
+ size 306807976
epoch225/training_config.toml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ output_dir = "/tmp/job_2585f190-3bf9-4398-acc3-9030f40fcf0f/training_outputs/wan22_t2v_low_lora"
2
+ dataset = "/tmp/job_2585f190-3bf9-4398-acc3-9030f40fcf0f/dataset.toml"
3
+ epochs = 300
4
+ micro_batch_size_per_gpu = 1
5
+ pipeline_stages = 1
6
+ gradient_accumulation_steps = 1
7
+ gradient_clipping = 1.0
8
+ warmup_steps = 100
9
+ save_every_n_epochs = 25
10
+ checkpoint_every_n_minutes = 120
11
+ activation_checkpointing = true
12
+ partition_method = "parameters"
13
+ save_dtype = "bfloat16"
14
+ caching_batch_size = 1
15
+ steps_per_print = 1
16
+ video_clip_mode = "single_middle"
17
+
18
+ [model]
19
+ type = "wan"
20
+ ckpt_path = "/workspace/models/models/Wan/Wan2.2-T2V-A14B"
21
+ transformer_path = "/workspace/models/models/Wan/Wan2.2-T2V-A14B/low_noise_model"
22
+ dtype = "bfloat16"
23
+ transformer_dtype = "float8"
24
+ min_t = 0
25
+ max_t = 0.875
26
+ timestep_sample_method = "logit_normal"
27
+
28
+ [adapter]
29
+ type = "lora"
30
+ rank = 32
31
+ dtype = "bfloat16"
32
+
33
+ [optimizer]
34
+ type = "AdamW8bitKahan"
35
+ lr = 2e-5
36
+ betas = [ 0.9, 0.99,]
37
+ weight_decay = 0.01
38
+ eps = 1e-8