Training in progress, step 1600

.gitignore

@@ -0,0 +1 @@
+checkpoint-*/

added_tokens.json

@@ -0,0 +1,168 @@
+{
+  "[CLS0]": 50269,
+  "[CLS1]": 50265,
+  "[CLS2]": 50377,
+  "[CLSN]": 50257,
+  "[INIT]": 50258,
+  "[NOARG]": 50259,
+  "[v0]": 50409,
+  "[v100]": 50381,
+  "[v101]": 50389,
+  "[v102]": 50400,
+  "[v103]": 50309,
+  "[v104]": 50301,
+  "[v105]": 50348,
+  "[v106]": 50297,
+  "[v107]": 50310,
+  "[v108]": 50390,
+  "[v109]": 50338,
+  "[v10]": 50279,
+  "[v110]": 50290,
+  "[v111]": 50370,
+  "[v112]": 50282,
+  "[v113]": 50299,
+  "[v114]": 50283,
+  "[v115]": 50300,
+  "[v116]": 50422,
+  "[v117]": 50305,
+  "[v118]": 50277,
+  "[v119]": 50383,
+  "[v11]": 50342,
+  "[v120]": 50306,
+  "[v121]": 50376,
+  "[v122]": 50405,
+  "[v123]": 50355,
+  "[v124]": 50328,
+  "[v125]": 50275,
+  "[v126]": 50395,
+  "[v127]": 50359,
+  "[v128]": 50415,
+  "[v129]": 50264,
+  "[v12]": 50360,
+  "[v130]": 50352,
+  "[v131]": 50322,
+  "[v132]": 50261,
+  "[v133]": 50393,
+  "[v134]": 50354,
+  "[v135]": 50418,
+  "[v136]": 50285,
+  "[v137]": 50281,
+  "[v138]": 50407,
+  "[v139]": 50325,
+  "[v13]": 50336,
+  "[v140]": 50396,
+  "[v141]": 50375,
+  "[v142]": 50372,
+  "[v143]": 50403,
+  "[v144]": 50378,
+  "[v145]": 50270,
+  "[v146]": 50339,
+  "[v147]": 50262,
+  "[v148]": 50399,
+  "[v149]": 50289,
+  "[v14]": 50385,
+  "[v150]": 50321,
+  "[v151]": 50404,
+  "[v152]": 50308,
+  "[v153]": 50368,
+  "[v154]": 50362,
+  "[v155]": 50412,
+  "[v156]": 50346,
+  "[v157]": 50260,
+  "[v158]": 50416,
+  "[v159]": 50294,
+  "[v15]": 50380,
+  "[v16]": 50334,
+  "[v17]": 50373,
+  "[v18]": 50365,
+  "[v19]": 50268,
+  "[v1]": 50408,
+  "[v20]": 50311,
+  "[v21]": 50401,
+  "[v22]": 50271,
+  "[v23]": 50361,
+  "[v24]": 50266,
+  "[v25]": 50291,
+  "[v26]": 50320,
+  "[v27]": 50329,
+  "[v28]": 50417,
+  "[v29]": 50356,
+  "[v2]": 50302,
+  "[v30]": 50347,
+  "[v31]": 50343,
+  "[v32]": 50421,
+  "[v33]": 50384,
+  "[v34]": 50280,
+  "[v35]": 50349,
+  "[v36]": 50330,
+  "[v37]": 50296,
+  "[v38]": 50357,
+  "[v39]": 50344,
+  "[v3]": 50284,
+  "[v40]": 50312,
+  "[v41]": 50298,
+  "[v42]": 50358,
+  "[v43]": 50276,
+  "[v44]": 50374,
+  "[v45]": 50263,
+  "[v46]": 50307,
+  "[v47]": 50313,
+  "[v48]": 50410,
+  "[v49]": 50324,
+  "[v4]": 50317,
+  "[v50]": 50272,
+  "[v51]": 50303,
+  "[v52]": 50392,
+  "[v53]": 50286,
+  "[v54]": 50341,
+  "[v55]": 50371,
+  "[v56]": 50367,
+  "[v57]": 50391,
+  "[v58]": 50379,
+  "[v59]": 50351,
+  "[v5]": 50304,
+  "[v60]": 50318,
+  "[v61]": 50394,
+  "[v62]": 50331,
+  "[v63]": 50337,
+  "[v64]": 50333,
+  "[v65]": 50278,
+  "[v66]": 50273,
+  "[v67]": 50326,
+  "[v68]": 50274,
+  "[v69]": 50386,
+  "[v6]": 50319,
+  "[v70]": 50335,
+  "[v71]": 50402,
+  "[v72]": 50292,
+  "[v73]": 50345,
+  "[v74]": 50293,
+  "[v75]": 50363,
+  "[v76]": 50397,
+  "[v77]": 50419,
+  "[v78]": 50382,
+  "[v79]": 50398,
+  "[v7]": 50267,
+  "[v80]": 50323,
+  "[v81]": 50288,
+  "[v82]": 50353,
+  "[v83]": 50364,
+  "[v84]": 50332,
+  "[v85]": 50387,
+  "[v86]": 50411,
+  "[v87]": 50413,
+  "[v88]": 50315,
+  "[v89]": 50366,
+  "[v8]": 50314,
+  "[v90]": 50406,
+  "[v91]": 50295,
+  "[v92]": 50340,
+  "[v93]": 50287,
+  "[v94]": 50388,
+  "[v95]": 50316,
+  "[v96]": 50414,
+  "[v97]": 50350,
+  "[v98]": 50369,
+  "[v99]": 50327,
+  "[v9]": 50420
+}

config.json

@@ -0,0 +1,47 @@
+{
+  "_name_or_path": "distilgpt2",
+  "_num_labels": 1,
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "id2label": {
+    "0": "LABEL_0"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "LABEL_0": 0
+  },
+  "layer_norm_epsilon": 1e-05,
+  "max_length": 912,
+  "model_type": "gpt2",
+  "n_ctx": 912,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 6,
+  "n_positions": 1024,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.24.0",
+  "use_cache": true,
+  "vocab_size": 50423
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5078e8045ccc0de0404391563da1d9c61ea38afa94f93efa6590328237b64099
+size 334480121

special_tokens_map.json

@@ -0,0 +1,174 @@
+{
+  "additional_special_tokens": [
+    "[CLSN]",
+    "[INIT]",
+    "[NOARG]",
+    "[v157]",
+    "[v132]",
+    "[v147]",
+    "[v45]",
+    "[v129]",
+    "[CLS1]",
+    "[v24]",
+    "[v7]",
+    "[v19]",
+    "[CLS0]",
+    "[v145]",
+    "[v22]",
+    "[v50]",
+    "[v66]",
+    "[v68]",
+    "[v125]",
+    "[v43]",
+    "[v118]",
+    "[v65]",
+    "[v10]",
+    "[v34]",
+    "[v137]",
+    "[v112]",
+    "[v114]",
+    "[v3]",
+    "[v136]",
+    "[v53]",
+    "[v93]",
+    "[v81]",
+    "[v149]",
+    "[v110]",
+    "[v25]",
+    "[v72]",
+    "[v74]",
+    "[v159]",
+    "[v91]",
+    "[v37]",
+    "[v106]",
+    "[v41]",
+    "[v113]",
+    "[v115]",
+    "[v104]",
+    "[v2]",
+    "[v51]",
+    "[v5]",
+    "[v117]",
+    "[v120]",
+    "[v46]",
+    "[v152]",
+    "[v103]",
+    "[v107]",
+    "[v20]",
+    "[v40]",
+    "[v47]",
+    "[v8]",
+    "[v88]",
+    "[v95]",
+    "[v4]",
+    "[v60]",
+    "[v6]",
+    "[v26]",
+    "[v150]",
+    "[v131]",
+    "[v80]",
+    "[v49]",
+    "[v139]",
+    "[v67]",
+    "[v99]",
+    "[v124]",
+    "[v27]",
+    "[v36]",
+    "[v62]",
+    "[v84]",
+    "[v64]",
+    "[v16]",
+    "[v70]",
+    "[v13]",
+    "[v63]",
+    "[v109]",
+    "[v146]",
+    "[v92]",
+    "[v54]",
+    "[v11]",
+    "[v31]",
+    "[v39]",
+    "[v73]",
+    "[v156]",
+    "[v30]",
+    "[v105]",
+    "[v35]",
+    "[v97]",
+    "[v59]",
+    "[v130]",
+    "[v82]",
+    "[v134]",
+    "[v123]",
+    "[v29]",
+    "[v38]",
+    "[v42]",
+    "[v127]",
+    "[v12]",
+    "[v23]",
+    "[v154]",
+    "[v75]",
+    "[v83]",
+    "[v18]",
+    "[v89]",
+    "[v56]",
+    "[v153]",
+    "[v98]",
+    "[v111]",
+    "[v55]",
+    "[v142]",
+    "[v17]",
+    "[v44]",
+    "[v141]",
+    "[v121]",
+    "[CLS2]",
+    "[v144]",
+    "[v58]",
+    "[v15]",
+    "[v100]",
+    "[v78]",
+    "[v119]",
+    "[v33]",
+    "[v14]",
+    "[v69]",
+    "[v85]",
+    "[v94]",
+    "[v101]",
+    "[v108]",
+    "[v57]",
+    "[v52]",
+    "[v133]",
+    "[v61]",
+    "[v126]",
+    "[v140]",
+    "[v76]",
+    "[v79]",
+    "[v148]",
+    "[v102]",
+    "[v21]",
+    "[v71]",
+    "[v143]",
+    "[v151]",
+    "[v122]",
+    "[v90]",
+    "[v138]",
+    "[v1]",
+    "[v0]",
+    "[v48]",
+    "[v86]",
+    "[v155]",
+    "[v87]",
+    "[v96]",
+    "[v128]",
+    "[v158]",
+    "[v28]",
+    "[v135]",
+    "[v77]",
+    "[v9]",
+    "[v32]",
+    "[v116]"
+  ],
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,10 @@
+{
+  "add_prefix_space": false,
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1024,
+  "name_or_path": "distilgpt2",
+  "special_tokens_map_file": null,
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8df4cea213788e01c2df8a4e8f049e1ac020ea1ad66ce9260293cc141e9af6ee
+size 3387