Add GPT-2 Medium Mixed FT for exp_loo_without_DM_Mathematics
Browse files- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/config.json +43 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/generation_config.json +6 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/model.safetensors +3 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/source_dataset_stats.csv +10 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/tokenizer.json +0 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/tokenizer_config.json +13 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/training_history.csv +3 -0
- gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/training_summary.csv +2 -0
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_function": "gelu_new",
|
| 3 |
+
"add_cross_attention": false,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"dtype": "float32",
|
| 10 |
+
"embd_pdrop": 0.1,
|
| 11 |
+
"eos_token_id": 50256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"layer_norm_epsilon": 1e-05,
|
| 14 |
+
"model_type": "gpt2",
|
| 15 |
+
"n_ctx": 1024,
|
| 16 |
+
"n_embd": 1024,
|
| 17 |
+
"n_head": 16,
|
| 18 |
+
"n_inner": null,
|
| 19 |
+
"n_layer": 24,
|
| 20 |
+
"n_positions": 1024,
|
| 21 |
+
"n_special": 0,
|
| 22 |
+
"pad_token_id": null,
|
| 23 |
+
"predict_special_tokens": true,
|
| 24 |
+
"reorder_and_upcast_attn": false,
|
| 25 |
+
"resid_pdrop": 0.1,
|
| 26 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 27 |
+
"scale_attn_weights": true,
|
| 28 |
+
"summary_activation": null,
|
| 29 |
+
"summary_first_dropout": 0.1,
|
| 30 |
+
"summary_proj_to_labels": true,
|
| 31 |
+
"summary_type": "cls_index",
|
| 32 |
+
"summary_use_proj": true,
|
| 33 |
+
"task_specific_params": {
|
| 34 |
+
"text-generation": {
|
| 35 |
+
"do_sample": true,
|
| 36 |
+
"max_length": 50
|
| 37 |
+
}
|
| 38 |
+
},
|
| 39 |
+
"tie_word_embeddings": true,
|
| 40 |
+
"transformers_version": "5.14.1",
|
| 41 |
+
"use_cache": true,
|
| 42 |
+
"vocab_size": 50257
|
| 43 |
+
}
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "5.14.1"
|
| 6 |
+
}
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e47297289b0ae687badf2e91cb454ccb0d53d51dc32930c3c5d3faef7f3ad10
|
| 3 |
+
size 1419322880
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/source_dataset_stats.csv
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
num_documents_used,raw_tokens_seen,requested_token_budget,usable_token_budget,num_sequences,context_length,effective_tokens,source,final_mixture_tokens
|
| 2 |
+
33,580056,569344,569344,1112,512,569344,ArXiv,569344
|
| 3 |
+
18574,550276,568832,568832,1111,512,568832,EuroParl,568832
|
| 4 |
+
254,569039,568832,568832,1111,512,568832,GitHub,568832
|
| 5 |
+
1,1200631,568832,568832,1111,512,568832,PG-19,568832
|
| 6 |
+
573,568366,568832,568832,1111,512,568832,OWT2,568832
|
| 7 |
+
1965,566870,568832,568832,1111,512,568832,PubMed Abstracts,568832
|
| 8 |
+
73,578162,568832,568832,1111,512,568832,PubMed Central,568832
|
| 9 |
+
2792,566077,568832,568832,1111,512,568832,StackExchange,568832
|
| 10 |
+
82,570941,568832,568832,1111,512,568832,Wikipedia,568832
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<|endoftext|>",
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"errors": "replace",
|
| 7 |
+
"is_local": false,
|
| 8 |
+
"local_files_only": false,
|
| 9 |
+
"model_max_length": 1024,
|
| 10 |
+
"pad_token": "<|endoftext|>",
|
| 11 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 12 |
+
"unk_token": "<|endoftext|>"
|
| 13 |
+
}
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/training_history.csv
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch,optimizer_step,cumulative_tokens_seen,optimization_loss,train_loss,val_loss,train_ppl,val_ppl,train_acc,val_acc,train_eval_tokens,val_eval_tokens,learning_rate
|
| 2 |
+
1,563,4608000,2.6094508269627887,2.34769711572165,2.468527784856034,10.46145045366329,11.805054472156248,0.5236959167662367,0.5097283457289278,4577181,508514,5e-05
|
| 3 |
+
2,1126,9216000,2.420128697925144,2.20034732037955,2.4289776613512606,9.028148614961427,11.347275389839485,0.5436304572617949,0.5148452156676119,4577181,508514,5e-05
|
gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model/training_summary.csv
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
method,experiment_name,model,model_name,hf_repo_id,hf_subfolder,sources,num_sources,mixture_type,seed,context_length,requested_total_max_tokens,effective_total_source_tokens,batch_size,gradient_accumulation_steps,effective_batch_size,learning_rate,weight_decay,val_split_ratio,max_epochs,final_optimizer_step,final_tokens_seen,final_train_loss,final_val_loss
|
| 2 |
+
Mixed FT,exp_loo_without_DM_Mathematics,medium,gpt2-medium,alignment-decision-lab/robustness-model-bank,gpt2Medium/Mixed_FT_ArXiv_EuroParl_GitHub_PG-19_OWT2_PubMed_Abstracts_PubMed_Central_StackExchange_Wikipedia/model,ArXiv | EuroParl | GitHub | PG-19 | OWT2 | PubMed Abstracts | PubMed Central | StackExchange | Wikipedia,9,uniform_source_mixture,42,512,5120000,5120000,8,2,16,5e-05,0.0,0.1,2,1126,9216000,2.20034732037955,2.4289776613512606
|