Upload configuration file
Browse files
config/fig1_all_openthoughts.yaml
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
operators:
|
| 2 |
+
- id: load_hf
|
| 3 |
+
config:
|
| 4 |
+
type: hf_source
|
| 5 |
+
dataset: open-thoughts/OpenThoughts-114k
|
| 6 |
+
split: train
|
| 7 |
+
subset: metadata
|
| 8 |
+
trust_remote_code: True
|
| 9 |
+
- id: decontaminate
|
| 10 |
+
config:
|
| 11 |
+
type: cpu_function
|
| 12 |
+
sharded: false
|
| 13 |
+
num_cpus: 16
|
| 14 |
+
function: data_strategies._A1_Source.utils.decontaminate_fuzzy_w_ngram
|
| 15 |
+
function_config:
|
| 16 |
+
column: problem
|
| 17 |
+
eval_datasets:
|
| 18 |
+
- HuggingFaceH4/MATH-500
|
| 19 |
+
- Maxwell-Jia/AIME_2024
|
| 20 |
+
- AI-MO/aimo-validation-amc
|
| 21 |
+
- livecodebench/code_generation_lite
|
| 22 |
+
- mlfoundations-dev/AIME2025_combined
|
| 23 |
+
- cais/hle
|
| 24 |
+
- open-r1/codeforces
|
| 25 |
+
- Idavidrein/gpqa
|
| 26 |
+
- daman1209arora/jeebench
|
| 27 |
+
- mlfoundations-dev/mmlu_pro_eval_full
|
| 28 |
+
- Qwen/CodeElo
|
| 29 |
+
- open-r1/ioi
|
| 30 |
+
eval_columns:
|
| 31 |
+
- problem
|
| 32 |
+
- Problem
|
| 33 |
+
- problem
|
| 34 |
+
- question_content
|
| 35 |
+
- question
|
| 36 |
+
- question
|
| 37 |
+
- description
|
| 38 |
+
- Question
|
| 39 |
+
- question
|
| 40 |
+
- prompt
|
| 41 |
+
- description
|
| 42 |
+
- statement
|
| 43 |
+
eval_splits:
|
| 44 |
+
- test
|
| 45 |
+
- train
|
| 46 |
+
- train
|
| 47 |
+
- test
|
| 48 |
+
- train
|
| 49 |
+
- test
|
| 50 |
+
- test
|
| 51 |
+
- train
|
| 52 |
+
- test
|
| 53 |
+
- test
|
| 54 |
+
- test
|
| 55 |
+
- test
|
| 56 |
+
eval_subsets:
|
| 57 |
+
Idavidrein/gpqa: gpqa_diamond
|
| 58 |
+
similarity_threshold: 75.0
|
| 59 |
+
ngram_size: 13
|
| 60 |
+
input_ids:
|
| 61 |
+
- load_hf
|
| 62 |
+
- id: convert_reasoning_trace_to_final
|
| 63 |
+
config:
|
| 64 |
+
type: function
|
| 65 |
+
function: data_strategies.commons.convert_reasoning_trace_to_final
|
| 66 |
+
function_config:
|
| 67 |
+
reasoning_column: deepseek_reasoning
|
| 68 |
+
solution_column: deepseek_solution
|
| 69 |
+
output_column: final_reasoning_trace
|
| 70 |
+
input_ids:
|
| 71 |
+
- decontaminate
|
| 72 |
+
- id: convert_to_sharegpt
|
| 73 |
+
config:
|
| 74 |
+
type: function
|
| 75 |
+
function: data_strategies.commons.convert_instruction_response_to_sharegpt
|
| 76 |
+
function_config:
|
| 77 |
+
input_instruction_column: problem
|
| 78 |
+
input_response_column: final_reasoning_trace
|
| 79 |
+
output_sharegpt_column: conversations
|
| 80 |
+
input_ids:
|
| 81 |
+
- convert_reasoning_trace_to_final
|