squash

README.md

@@ -1,3 +1,43 @@
 ---
 license: mit
+pipeline_tag: text-generation
+widget:
+- text: "@@ПЕРВЫЙ@@ привет @@ВТОРОЙ@@ привет @@ПЕРВЫЙ@@ как дела? @@ВТОРОЙ@@"
+  example_title: "how r u"
+- text: "@@ПЕРВЫЙ@@ что ты делал на выходных? @@ВТОРОЙ@@"
+  example_title: "wyd"
+language:
+- ru
+tags:
+- conversational
 ---
+This generation model is based on [sberbank-ai/rugpt3small_based_on_gpt2](https://huggingface.co/sberbank-ai/rugpt3small_based_on_gpt2). It's trained on large corpus of dialog data and can be used for buildning generative conversational agents
+
+The model was trained with context size 3
+
+
+On a validation set we calculated metrics introduced in [this paper](https://arxiv.org/pdf/2001.09977.pdf): 
+- Sensibleness: Operators are asked whether model's response makes sense given the context
+- Specificity: Operators are asked whether model's response is specific for given context, in other words we don't want our model to give general and boring responses
+- SSA which is the average of two metrics above (Sensibleness Specificity Average)
+
+|                                                     |   sensibleness |   specificity |   SSA |
+|:----------------------------------------------------|---------------:|--------------:|------:|
+| [tinkoff-ai/ruDialoGPT-small](https://huggingface.co/tinkoff-ai/ruDialoGPT-small)  |           0.64 |          0.5  | 0.57  |
+| [tinkoff-ai/ruDialoGPT-medium](https://huggingface.co/tinkoff-ai/ruDialoGPT-medium) |           0.78 |          0.69 | 0.735 |
+
+
+How to use:
+
+```python
+import torch
+from transformers import AutoTokenizer, AutoModelWithLMHead
+
+tokenizer = AutoTokenizer.from_pretrained('tinkoff-ai/ruDialoGPT-small')
+model = AutoModelWithLMHead.from_pretrained('tinkoff-ai/ruDialoGPT-small')
+inputs = tokenizer('@@ПЕРВЫЙ@@ привет @@ВТОРОЙ@@ привет @@ПЕРВЫЙ@@ как дела? @@ВТОРОЙ@@', return_tensors='pt')
+with torch.inference_mode():
+    generated_token_ids = model.generate(**inputs)
+    context_with_response = tokenizer.decode(generated_token_ids[0])
+context_with_response
+```

added_tokens.json

@@ -0,0 +1 @@
+{"@@ПЕРВЫЙ@@": 50257, "@@ВТОРОЙ@@": 50258, "<FIRST_SPEAKER>": 50259, "<SECOND_SPEAKER>": 50260}

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 3.6833789348602295,
+    "eval_runtime": 180.5086,
+    "eval_samples": 40311,
+    "eval_samples_per_second": 223.319,
+    "eval_steps_per_second": 18.614,
+    "perplexity": 39.780583236136096,
+    "train_loss": 3.773873895684371,
+    "train_runtime": 142283.9744,
+    "train_samples": 8570756,
+    "train_samples_per_second": 60.237,
+    "train_steps_per_second": 5.02
+}

config.json

@@ -0,0 +1,30 @@
+{
+  "activation_function": "gelu_new",
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "gradient_checkpointing": false,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 2048,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 2048,
+  "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,
+  "torch_dtype": "float32",
+  "transformers_version": "4.15.0",
+  "use_cache": true,
+  "vocab_size": 50261
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 3.6833789348602295,
+    "eval_runtime": 180.5086,
+    "eval_samples": 40311,
+    "eval_samples_per_second": 223.319,
+    "eval_steps_per_second": 18.614,
+    "perplexity": 39.780583236136096
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf9455a4326b0038ec5f80baca7f49498f863459a1e816de8ce9096edd2efb9c
+size 551310569

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "pad_token": "<pad>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"errors": "replace", "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "special_tokens_map_file": null, "use_fast": true, "tokenizer_class": "GPT2Tokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "train_loss": 3.773873895684371,
+    "train_runtime": 142283.9744,
+    "train_samples": 8570756,
+    "train_samples_per_second": 60.237,
+    "train_steps_per_second": 5.02
+}