sparkleman commited on
Commit
684d4c2
·
1 Parent(s): 4350c55

UPDATE: RWKV G1 2.9B 32%

Browse files
config.production-modelscope.yaml CHANGED
@@ -4,8 +4,8 @@ STRATEGY: "cuda fp16"
4
  RWKV_CUDA_ON: True
5
  CHUNK_LEN: 256
6
  MODELS:
7
- - SERVICE_NAME: "RWKV7-G1-2.9B-16%trained-20250313-ctx4k"
8
- DOWNLOAD_MODEL_FILE_NAME: "RWKV7-G1-2.9B-16%trained-20250313-ctx4k.pth"
9
  DOWNLOAD_MODEL_REPO_ID: "RWKV/temp-latest-training-models"
10
  REASONING: True
11
  DEFAULT_SAMPLER:
 
4
  RWKV_CUDA_ON: True
5
  CHUNK_LEN: 256
6
  MODELS:
7
+ - SERVICE_NAME: "RWKV7-G1-2.9B-32%trained-20250327-ctx4k"
8
+ DOWNLOAD_MODEL_FILE_NAME: "RWKV7-G1-2.9B-32%trained-20250327-ctx4k.pth"
9
  DOWNLOAD_MODEL_REPO_ID: "RWKV/temp-latest-training-models"
10
  REASONING: True
11
  DEFAULT_SAMPLER:
config.production.yaml CHANGED
@@ -4,8 +4,8 @@ STRATEGY: "cuda fp16"
4
  RWKV_CUDA_ON: True
5
  CHUNK_LEN: 256
6
  MODELS:
7
- - SERVICE_NAME: "RWKV7-G1-2.9B-16%trained-20250313-ctx4k"
8
- DOWNLOAD_MODEL_FILE_NAME: "RWKV7-G1-2.9B-16%trained-20250313-ctx4k.pth"
9
  DOWNLOAD_MODEL_REPO_ID: "BlinkDL/temp-latest-training-models"
10
  REASONING: True
11
  DEFAULT_SAMPLER:
 
4
  RWKV_CUDA_ON: True
5
  CHUNK_LEN: 256
6
  MODELS:
7
+ - SERVICE_NAME: "RWKV7-G1-2.9B-32%trained-20250327-ctx4k"
8
+ DOWNLOAD_MODEL_FILE_NAME: "RWKV7-G1-2.9B-32%trained-20250327-ctx4k.pth"
9
  DOWNLOAD_MODEL_REPO_ID: "BlinkDL/temp-latest-training-models"
10
  REASONING: True
11
  DEFAULT_SAMPLER: