mnoukhov/pythia410m-sft-tldr
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: mnoukhov/pythia410m-sft-tldr
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-30 14:21:21 +08:00
commit 4bb3a5553f
215 changed files with 118927 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
code/configs/vllm_20k_costa_1b_dpo.yml Normal file
View File

@@ -0,0 +1,14 @@
output_dir: summarize_from_feedback_tldr3_generated_relabel_20k_dpo_costa_1b_fp16.yml_3d94f50_b35a8
dataset_name: mnoukhov/summarize_from_feedback_tldr_3_filtered_oai_preprocessing_1706381144_unlabelled
model_name: mnoukhov/EleutherAI_pythia-1b-deduped__sft__tldr_dpo_costa_1b_fp16.yml_3d94f50_b35a8
base_model_name: vwxyzjn/EleutherAI_pythia-1b-deduped__sft__tldr
base_model_revision: sft__55513__1706646024
lora_model: True
tokenizer_name: EleutherAI/pythia-1b-deduped
train_split: train[:20000]
batch_size: 4
max_prompt_length: 512
max_new_tokens: 128
temperature: 0.7
top_p: 1.0
dtype: bfloat16