upload checkpoint-175
Browse files- .gitattributes +1 -0
- adapter_config.json +48 -0
- adapter_model.safetensors +3 -0
- chat_template.jinja +61 -0
- tokenizer.json +3 -0
- tokenizer_config.json +30 -0
- trainer_state.json +2483 -0
- training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
adapter_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alora_invocation_tokens": null,
|
| 3 |
+
"alpha_pattern": {},
|
| 4 |
+
"arrow_config": null,
|
| 5 |
+
"auto_mapping": null,
|
| 6 |
+
"base_model_name_or_path": "Qwen/Qwen3-30B-A3B-Instruct-2507",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"corda_config": null,
|
| 9 |
+
"ensure_weight_tying": false,
|
| 10 |
+
"eva_config": null,
|
| 11 |
+
"exclude_modules": null,
|
| 12 |
+
"fan_in_fan_out": null,
|
| 13 |
+
"inference_mode": true,
|
| 14 |
+
"init_lora_weights": true,
|
| 15 |
+
"layer_replication": null,
|
| 16 |
+
"layers_pattern": null,
|
| 17 |
+
"layers_to_transform": null,
|
| 18 |
+
"loftq_config": {},
|
| 19 |
+
"lora_alpha": 32,
|
| 20 |
+
"lora_bias": false,
|
| 21 |
+
"lora_dropout": 0.0,
|
| 22 |
+
"lora_ga_config": null,
|
| 23 |
+
"megatron_config": null,
|
| 24 |
+
"megatron_core": "megatron.core",
|
| 25 |
+
"modules_to_save": null,
|
| 26 |
+
"peft_type": "LORA",
|
| 27 |
+
"peft_version": "0.19.1",
|
| 28 |
+
"qalora_group_size": 16,
|
| 29 |
+
"r": 16,
|
| 30 |
+
"rank_pattern": {},
|
| 31 |
+
"revision": null,
|
| 32 |
+
"target_modules": [
|
| 33 |
+
"q_proj",
|
| 34 |
+
"k_proj",
|
| 35 |
+
"v_proj",
|
| 36 |
+
"o_proj"
|
| 37 |
+
],
|
| 38 |
+
"target_parameters": [
|
| 39 |
+
"mlp.experts.gate_up_proj",
|
| 40 |
+
"mlp.experts.down_proj"
|
| 41 |
+
],
|
| 42 |
+
"task_type": "CAUSAL_LM",
|
| 43 |
+
"trainable_token_indices": null,
|
| 44 |
+
"use_bdlora": null,
|
| 45 |
+
"use_dora": false,
|
| 46 |
+
"use_qalora": false,
|
| 47 |
+
"use_rslora": false
|
| 48 |
+
}
|
adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3208620f28b21da37e985065cd88f8d8cf832cacfdd591389f512203cad18d4f
|
| 3 |
+
size 2570139624
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0].role == 'system' %}
|
| 4 |
+
{{- messages[0].content + '\n\n' }}
|
| 5 |
+
{%- endif %}
|
| 6 |
+
{{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 7 |
+
{%- for tool in tools %}
|
| 8 |
+
{{- "\n" }}
|
| 9 |
+
{{- tool | tojson }}
|
| 10 |
+
{%- endfor %}
|
| 11 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 12 |
+
{%- else %}
|
| 13 |
+
{%- if messages[0].role == 'system' %}
|
| 14 |
+
{{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
|
| 15 |
+
{%- endif %}
|
| 16 |
+
{%- endif %}
|
| 17 |
+
{%- for message in messages %}
|
| 18 |
+
{%- if message.content is string %}
|
| 19 |
+
{%- set content = message.content %}
|
| 20 |
+
{%- else %}
|
| 21 |
+
{%- set content = '' %}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
|
| 24 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 25 |
+
{%- elif message.role == "assistant" %}
|
| 26 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 27 |
+
{%- if message.tool_calls %}
|
| 28 |
+
{%- for tool_call in message.tool_calls %}
|
| 29 |
+
{%- if (loop.first and content) or (not loop.first) %}
|
| 30 |
+
{{- '\n' }}
|
| 31 |
+
{%- endif %}
|
| 32 |
+
{%- if tool_call.function %}
|
| 33 |
+
{%- set tool_call = tool_call.function %}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{{- '<tool_call>\n{"name": "' }}
|
| 36 |
+
{{- tool_call.name }}
|
| 37 |
+
{{- '", "arguments": ' }}
|
| 38 |
+
{%- if tool_call.arguments is string %}
|
| 39 |
+
{{- tool_call.arguments }}
|
| 40 |
+
{%- else %}
|
| 41 |
+
{{- tool_call.arguments | tojson }}
|
| 42 |
+
{%- endif %}
|
| 43 |
+
{{- '}\n</tool_call>' }}
|
| 44 |
+
{%- endfor %}
|
| 45 |
+
{%- endif %}
|
| 46 |
+
{{- '<|im_end|>\n' }}
|
| 47 |
+
{%- elif message.role == "tool" %}
|
| 48 |
+
{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
|
| 49 |
+
{{- '<|im_start|>user' }}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{{- '\n<tool_response>\n' }}
|
| 52 |
+
{{- content }}
|
| 53 |
+
{{- '\n</tool_response>' }}
|
| 54 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 55 |
+
{{- '<|im_end|>\n' }}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- endif %}
|
| 58 |
+
{%- endfor %}
|
| 59 |
+
{%- if add_generation_prompt %}
|
| 60 |
+
{{- '<|im_start|>assistant\n' }}
|
| 61 |
+
{%- endif %}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
|
| 3 |
+
size 11422650
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": null,
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<|im_end|>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"<|im_start|>",
|
| 10 |
+
"<|im_end|>",
|
| 11 |
+
"<|object_ref_start|>",
|
| 12 |
+
"<|object_ref_end|>",
|
| 13 |
+
"<|box_start|>",
|
| 14 |
+
"<|box_end|>",
|
| 15 |
+
"<|quad_start|>",
|
| 16 |
+
"<|quad_end|>",
|
| 17 |
+
"<|vision_start|>",
|
| 18 |
+
"<|vision_end|>",
|
| 19 |
+
"<|vision_pad|>",
|
| 20 |
+
"<|image_pad|>",
|
| 21 |
+
"<|video_pad|>"
|
| 22 |
+
],
|
| 23 |
+
"is_local": false,
|
| 24 |
+
"local_files_only": false,
|
| 25 |
+
"model_max_length": 1010000,
|
| 26 |
+
"pad_token": "<|endoftext|>",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null
|
| 30 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2483 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.23248090335436733,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 175,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.001328462304882099,
|
| 14 |
+
"grad_norm": 4.078240394592285,
|
| 15 |
+
"learning_rate": 0.0,
|
| 16 |
+
"loss": 1.2204899787902832,
|
| 17 |
+
"memory/device_reserved (GiB)": 74.0,
|
| 18 |
+
"memory/max_active (GiB)": 73.31,
|
| 19 |
+
"memory/max_allocated (GiB)": 73.31,
|
| 20 |
+
"ppl": 3.38885,
|
| 21 |
+
"step": 1,
|
| 22 |
+
"tokens/total": 241280,
|
| 23 |
+
"tokens/trainable": 1844
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.002656924609764198,
|
| 27 |
+
"grad_norm": 3.4430620670318604,
|
| 28 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 29 |
+
"loss": 1.208996057510376,
|
| 30 |
+
"memory/device_reserved (GiB)": 74.01,
|
| 31 |
+
"memory/max_active (GiB)": 71.94,
|
| 32 |
+
"memory/max_allocated (GiB)": 71.94,
|
| 33 |
+
"ppl": 3.35012,
|
| 34 |
+
"step": 2,
|
| 35 |
+
"tokens/total": 481728,
|
| 36 |
+
"tokens/train_per_sec_per_gpu": 13.18,
|
| 37 |
+
"tokens/trainable": 3740
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.003985386914646297,
|
| 41 |
+
"grad_norm": 3.8220441341400146,
|
| 42 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 43 |
+
"loss": 1.428513526916504,
|
| 44 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 45 |
+
"memory/max_active (GiB)": 75.91,
|
| 46 |
+
"memory/max_allocated (GiB)": 75.91,
|
| 47 |
+
"ppl": 4.17249,
|
| 48 |
+
"step": 3,
|
| 49 |
+
"tokens/total": 699776,
|
| 50 |
+
"tokens/train_per_sec_per_gpu": 13.14,
|
| 51 |
+
"tokens/trainable": 5674
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.005313849219528396,
|
| 55 |
+
"grad_norm": 3.1648757457733154,
|
| 56 |
+
"learning_rate": 6.818181818181818e-06,
|
| 57 |
+
"loss": 1.1087837219238281,
|
| 58 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 59 |
+
"memory/max_active (GiB)": 75.18,
|
| 60 |
+
"memory/max_allocated (GiB)": 75.18,
|
| 61 |
+
"ppl": 3.03067,
|
| 62 |
+
"step": 4,
|
| 63 |
+
"tokens/total": 961536,
|
| 64 |
+
"tokens/train_per_sec_per_gpu": 10.39,
|
| 65 |
+
"tokens/trainable": 7547
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.006642311524410495,
|
| 69 |
+
"grad_norm": 3.243049144744873,
|
| 70 |
+
"learning_rate": 9.090909090909091e-06,
|
| 71 |
+
"loss": 1.2332323789596558,
|
| 72 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 73 |
+
"memory/max_active (GiB)": 75.5,
|
| 74 |
+
"memory/max_allocated (GiB)": 75.5,
|
| 75 |
+
"ppl": 3.43231,
|
| 76 |
+
"step": 5,
|
| 77 |
+
"tokens/total": 1220800,
|
| 78 |
+
"tokens/train_per_sec_per_gpu": 10.89,
|
| 79 |
+
"tokens/trainable": 9461
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.007970773829292593,
|
| 83 |
+
"grad_norm": 2.992734670639038,
|
| 84 |
+
"learning_rate": 1.1363636363636365e-05,
|
| 85 |
+
"loss": 1.1764527559280396,
|
| 86 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 87 |
+
"memory/max_active (GiB)": 75.28,
|
| 88 |
+
"memory/max_allocated (GiB)": 75.28,
|
| 89 |
+
"ppl": 3.24285,
|
| 90 |
+
"step": 6,
|
| 91 |
+
"tokens/total": 1486464,
|
| 92 |
+
"tokens/train_per_sec_per_gpu": 14.72,
|
| 93 |
+
"tokens/trainable": 12044
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.009299236134174693,
|
| 97 |
+
"grad_norm": 7.225315093994141,
|
| 98 |
+
"learning_rate": 1.3636363636363637e-05,
|
| 99 |
+
"loss": 1.3602099418640137,
|
| 100 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 101 |
+
"memory/max_active (GiB)": 74.3,
|
| 102 |
+
"memory/max_allocated (GiB)": 74.3,
|
| 103 |
+
"ppl": 3.89701,
|
| 104 |
+
"step": 7,
|
| 105 |
+
"tokens/total": 1721408,
|
| 106 |
+
"tokens/train_per_sec_per_gpu": 12.8,
|
| 107 |
+
"tokens/trainable": 14012
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.010627698439056792,
|
| 111 |
+
"grad_norm": 3.7612006664276123,
|
| 112 |
+
"learning_rate": 1.590909090909091e-05,
|
| 113 |
+
"loss": 1.3938989639282227,
|
| 114 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 115 |
+
"memory/max_active (GiB)": 74.33,
|
| 116 |
+
"memory/max_allocated (GiB)": 74.33,
|
| 117 |
+
"ppl": 4.03053,
|
| 118 |
+
"step": 8,
|
| 119 |
+
"tokens/total": 1929152,
|
| 120 |
+
"tokens/train_per_sec_per_gpu": 14.44,
|
| 121 |
+
"tokens/trainable": 15860
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.01195616074393889,
|
| 125 |
+
"grad_norm": 3.4705631732940674,
|
| 126 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 127 |
+
"loss": 1.2233916521072388,
|
| 128 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 129 |
+
"memory/max_active (GiB)": 72.51,
|
| 130 |
+
"memory/max_allocated (GiB)": 72.51,
|
| 131 |
+
"ppl": 3.3987,
|
| 132 |
+
"step": 9,
|
| 133 |
+
"tokens/total": 2134464,
|
| 134 |
+
"tokens/train_per_sec_per_gpu": 13.58,
|
| 135 |
+
"tokens/trainable": 17682
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.01328462304882099,
|
| 139 |
+
"grad_norm": 3.247605562210083,
|
| 140 |
+
"learning_rate": 2.0454545454545457e-05,
|
| 141 |
+
"loss": 1.082282543182373,
|
| 142 |
+
"memory/device_reserved (GiB)": 76.47,
|
| 143 |
+
"memory/max_active (GiB)": 72.07,
|
| 144 |
+
"memory/max_allocated (GiB)": 72.07,
|
| 145 |
+
"ppl": 2.95141,
|
| 146 |
+
"step": 10,
|
| 147 |
+
"tokens/total": 2366720,
|
| 148 |
+
"tokens/train_per_sec_per_gpu": 13.6,
|
| 149 |
+
"tokens/trainable": 19525
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.014613085353703089,
|
| 153 |
+
"grad_norm": 2.765240430831909,
|
| 154 |
+
"learning_rate": 2.272727272727273e-05,
|
| 155 |
+
"loss": 1.0681132078170776,
|
| 156 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 157 |
+
"memory/max_active (GiB)": 76.02,
|
| 158 |
+
"memory/max_allocated (GiB)": 76.02,
|
| 159 |
+
"ppl": 2.90988,
|
| 160 |
+
"step": 11,
|
| 161 |
+
"tokens/total": 2609600,
|
| 162 |
+
"tokens/train_per_sec_per_gpu": 12.68,
|
| 163 |
+
"tokens/trainable": 21534
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.015941547658585187,
|
| 167 |
+
"grad_norm": 2.2525298595428467,
|
| 168 |
+
"learning_rate": 2.5e-05,
|
| 169 |
+
"loss": 0.8673083782196045,
|
| 170 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 171 |
+
"memory/max_active (GiB)": 75.23,
|
| 172 |
+
"memory/max_allocated (GiB)": 75.23,
|
| 173 |
+
"ppl": 2.38049,
|
| 174 |
+
"step": 12,
|
| 175 |
+
"tokens/total": 2881216,
|
| 176 |
+
"tokens/train_per_sec_per_gpu": 11.34,
|
| 177 |
+
"tokens/trainable": 23688
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.017270009963467288,
|
| 181 |
+
"grad_norm": 2.2587997913360596,
|
| 182 |
+
"learning_rate": 2.7272727272727273e-05,
|
| 183 |
+
"loss": 0.8689092993736267,
|
| 184 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 185 |
+
"memory/max_active (GiB)": 75.53,
|
| 186 |
+
"memory/max_allocated (GiB)": 75.53,
|
| 187 |
+
"ppl": 2.38431,
|
| 188 |
+
"step": 13,
|
| 189 |
+
"tokens/total": 3154112,
|
| 190 |
+
"tokens/train_per_sec_per_gpu": 10.46,
|
| 191 |
+
"tokens/trainable": 25514
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.018598472268349386,
|
| 195 |
+
"grad_norm": 1.9934321641921997,
|
| 196 |
+
"learning_rate": 2.954545454545455e-05,
|
| 197 |
+
"loss": 0.7913857102394104,
|
| 198 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 199 |
+
"memory/max_active (GiB)": 74.19,
|
| 200 |
+
"memory/max_allocated (GiB)": 74.19,
|
| 201 |
+
"ppl": 2.20645,
|
| 202 |
+
"step": 14,
|
| 203 |
+
"tokens/total": 3410304,
|
| 204 |
+
"tokens/train_per_sec_per_gpu": 12.18,
|
| 205 |
+
"tokens/trainable": 27547
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.019926934573231483,
|
| 209 |
+
"grad_norm": 1.800851821899414,
|
| 210 |
+
"learning_rate": 3.181818181818182e-05,
|
| 211 |
+
"loss": 0.8316292762756348,
|
| 212 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 213 |
+
"memory/max_active (GiB)": 75.77,
|
| 214 |
+
"memory/max_allocated (GiB)": 75.77,
|
| 215 |
+
"ppl": 2.29706,
|
| 216 |
+
"step": 15,
|
| 217 |
+
"tokens/total": 3654464,
|
| 218 |
+
"tokens/train_per_sec_per_gpu": 13.47,
|
| 219 |
+
"tokens/trainable": 29471
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.021255396878113585,
|
| 223 |
+
"grad_norm": 1.455098271369934,
|
| 224 |
+
"learning_rate": 3.409090909090909e-05,
|
| 225 |
+
"loss": 0.6257059574127197,
|
| 226 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 227 |
+
"memory/max_active (GiB)": 75.2,
|
| 228 |
+
"memory/max_allocated (GiB)": 75.2,
|
| 229 |
+
"ppl": 1.86957,
|
| 230 |
+
"step": 16,
|
| 231 |
+
"tokens/total": 3909248,
|
| 232 |
+
"tokens/train_per_sec_per_gpu": 12.08,
|
| 233 |
+
"tokens/trainable": 31464
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.022583859182995682,
|
| 237 |
+
"grad_norm": 1.604706883430481,
|
| 238 |
+
"learning_rate": 3.6363636363636364e-05,
|
| 239 |
+
"loss": 0.7017816305160522,
|
| 240 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 241 |
+
"memory/max_active (GiB)": 74.85,
|
| 242 |
+
"memory/max_allocated (GiB)": 74.85,
|
| 243 |
+
"ppl": 2.01734,
|
| 244 |
+
"step": 17,
|
| 245 |
+
"tokens/total": 4131008,
|
| 246 |
+
"tokens/train_per_sec_per_gpu": 11.97,
|
| 247 |
+
"tokens/trainable": 33074
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.02391232148787778,
|
| 251 |
+
"grad_norm": 0.9868711233139038,
|
| 252 |
+
"learning_rate": 3.8636363636363636e-05,
|
| 253 |
+
"loss": 0.5830010175704956,
|
| 254 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 255 |
+
"memory/max_active (GiB)": 72.75,
|
| 256 |
+
"memory/max_allocated (GiB)": 72.75,
|
| 257 |
+
"ppl": 1.79141,
|
| 258 |
+
"step": 18,
|
| 259 |
+
"tokens/total": 4342272,
|
| 260 |
+
"tokens/train_per_sec_per_gpu": 15.53,
|
| 261 |
+
"tokens/trainable": 35185
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.02524078379275988,
|
| 265 |
+
"grad_norm": 0.8437089323997498,
|
| 266 |
+
"learning_rate": 4.0909090909090915e-05,
|
| 267 |
+
"loss": 0.49017488956451416,
|
| 268 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 269 |
+
"memory/max_active (GiB)": 75.23,
|
| 270 |
+
"memory/max_allocated (GiB)": 75.23,
|
| 271 |
+
"ppl": 1.6326,
|
| 272 |
+
"step": 19,
|
| 273 |
+
"tokens/total": 4594880,
|
| 274 |
+
"tokens/train_per_sec_per_gpu": 12.79,
|
| 275 |
+
"tokens/trainable": 37280
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.02656924609764198,
|
| 279 |
+
"grad_norm": 0.6587421298027039,
|
| 280 |
+
"learning_rate": 4.318181818181819e-05,
|
| 281 |
+
"loss": 0.3691103458404541,
|
| 282 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 283 |
+
"memory/max_active (GiB)": 75.34,
|
| 284 |
+
"memory/max_allocated (GiB)": 75.34,
|
| 285 |
+
"ppl": 1.44645,
|
| 286 |
+
"step": 20,
|
| 287 |
+
"tokens/total": 4858752,
|
| 288 |
+
"tokens/train_per_sec_per_gpu": 10.83,
|
| 289 |
+
"tokens/trainable": 39261
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.027897708402524077,
|
| 293 |
+
"grad_norm": 0.7513092160224915,
|
| 294 |
+
"learning_rate": 4.545454545454546e-05,
|
| 295 |
+
"loss": 0.4554915130138397,
|
| 296 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 297 |
+
"memory/max_active (GiB)": 74.66,
|
| 298 |
+
"memory/max_allocated (GiB)": 74.66,
|
| 299 |
+
"ppl": 1.57695,
|
| 300 |
+
"step": 21,
|
| 301 |
+
"tokens/total": 5107648,
|
| 302 |
+
"tokens/train_per_sec_per_gpu": 12.29,
|
| 303 |
+
"tokens/trainable": 41174
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.029226170707406178,
|
| 307 |
+
"grad_norm": 0.5032646059989929,
|
| 308 |
+
"learning_rate": 4.772727272727273e-05,
|
| 309 |
+
"loss": 0.4031963348388672,
|
| 310 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 311 |
+
"memory/max_active (GiB)": 72.94,
|
| 312 |
+
"memory/max_allocated (GiB)": 72.94,
|
| 313 |
+
"ppl": 1.4966,
|
| 314 |
+
"step": 22,
|
| 315 |
+
"tokens/total": 5310976,
|
| 316 |
+
"tokens/train_per_sec_per_gpu": 16.02,
|
| 317 |
+
"tokens/trainable": 43051
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.030554633012288276,
|
| 321 |
+
"grad_norm": 0.42180025577545166,
|
| 322 |
+
"learning_rate": 5e-05,
|
| 323 |
+
"loss": 0.44643938541412354,
|
| 324 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 325 |
+
"memory/max_active (GiB)": 76.1,
|
| 326 |
+
"memory/max_allocated (GiB)": 76.1,
|
| 327 |
+
"ppl": 1.56274,
|
| 328 |
+
"step": 23,
|
| 329 |
+
"tokens/total": 5557504,
|
| 330 |
+
"tokens/train_per_sec_per_gpu": 14.84,
|
| 331 |
+
"tokens/trainable": 45232
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.031883095317170373,
|
| 335 |
+
"grad_norm": 0.5073381662368774,
|
| 336 |
+
"learning_rate": 4.999976912636752e-05,
|
| 337 |
+
"loss": 0.31508195400238037,
|
| 338 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 339 |
+
"memory/max_active (GiB)": 75.94,
|
| 340 |
+
"memory/max_allocated (GiB)": 75.94,
|
| 341 |
+
"ppl": 1.37037,
|
| 342 |
+
"step": 24,
|
| 343 |
+
"tokens/total": 5795904,
|
| 344 |
+
"tokens/train_per_sec_per_gpu": 10.15,
|
| 345 |
+
"tokens/trainable": 46943
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.033211557622052475,
|
| 349 |
+
"grad_norm": 0.3897036910057068,
|
| 350 |
+
"learning_rate": 4.999907650973428e-05,
|
| 351 |
+
"loss": 0.45456668734550476,
|
| 352 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 353 |
+
"memory/max_active (GiB)": 74.17,
|
| 354 |
+
"memory/max_allocated (GiB)": 74.17,
|
| 355 |
+
"ppl": 1.57549,
|
| 356 |
+
"step": 25,
|
| 357 |
+
"tokens/total": 6040896,
|
| 358 |
+
"tokens/train_per_sec_per_gpu": 13.88,
|
| 359 |
+
"tokens/trainable": 48930
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.034540019926934576,
|
| 363 |
+
"grad_norm": 0.3590036630630493,
|
| 364 |
+
"learning_rate": 4.999792216289284e-05,
|
| 365 |
+
"loss": 0.3161405920982361,
|
| 366 |
+
"memory/device_reserved (GiB)": 77.06,
|
| 367 |
+
"memory/max_active (GiB)": 75.69,
|
| 368 |
+
"memory/max_allocated (GiB)": 75.69,
|
| 369 |
+
"ppl": 1.37182,
|
| 370 |
+
"step": 26,
|
| 371 |
+
"tokens/total": 6295680,
|
| 372 |
+
"tokens/train_per_sec_per_gpu": 9.24,
|
| 373 |
+
"tokens/trainable": 51085
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.03586848223181667,
|
| 377 |
+
"grad_norm": 0.45661401748657227,
|
| 378 |
+
"learning_rate": 4.999630610716385e-05,
|
| 379 |
+
"loss": 0.462736576795578,
|
| 380 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 381 |
+
"memory/max_active (GiB)": 76.65,
|
| 382 |
+
"memory/max_allocated (GiB)": 76.65,
|
| 383 |
+
"ppl": 1.58841,
|
| 384 |
+
"step": 27,
|
| 385 |
+
"tokens/total": 6534144,
|
| 386 |
+
"tokens/train_per_sec_per_gpu": 13.52,
|
| 387 |
+
"tokens/trainable": 53256
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.03719694453669877,
|
| 391 |
+
"grad_norm": 0.4172951579093933,
|
| 392 |
+
"learning_rate": 4.99942283723957e-05,
|
| 393 |
+
"loss": 0.45428770780563354,
|
| 394 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 395 |
+
"memory/max_active (GiB)": 73.51,
|
| 396 |
+
"memory/max_allocated (GiB)": 73.51,
|
| 397 |
+
"ppl": 1.57505,
|
| 398 |
+
"step": 28,
|
| 399 |
+
"tokens/total": 6745152,
|
| 400 |
+
"tokens/train_per_sec_per_gpu": 15.99,
|
| 401 |
+
"tokens/trainable": 55353
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.03852540684158087,
|
| 405 |
+
"grad_norm": 0.2786731421947479,
|
| 406 |
+
"learning_rate": 4.99916889969639e-05,
|
| 407 |
+
"loss": 0.32081466913223267,
|
| 408 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 409 |
+
"memory/max_active (GiB)": 75.75,
|
| 410 |
+
"memory/max_allocated (GiB)": 75.75,
|
| 411 |
+
"ppl": 1.37825,
|
| 412 |
+
"step": 29,
|
| 413 |
+
"tokens/total": 6994688,
|
| 414 |
+
"tokens/train_per_sec_per_gpu": 12.33,
|
| 415 |
+
"tokens/trainable": 57312
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.03985386914646297,
|
| 419 |
+
"grad_norm": 0.3878122866153717,
|
| 420 |
+
"learning_rate": 4.998868802777046e-05,
|
| 421 |
+
"loss": 0.43424832820892334,
|
| 422 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 423 |
+
"memory/max_active (GiB)": 76.4,
|
| 424 |
+
"memory/max_allocated (GiB)": 76.4,
|
| 425 |
+
"ppl": 1.5438,
|
| 426 |
+
"step": 30,
|
| 427 |
+
"tokens/total": 7220288,
|
| 428 |
+
"tokens/train_per_sec_per_gpu": 13.43,
|
| 429 |
+
"tokens/trainable": 59213
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.04118233145134507,
|
| 433 |
+
"grad_norm": 0.3197435140609741,
|
| 434 |
+
"learning_rate": 4.998522552024294e-05,
|
| 435 |
+
"loss": 0.40692442655563354,
|
| 436 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 437 |
+
"memory/max_active (GiB)": 76.43,
|
| 438 |
+
"memory/max_allocated (GiB)": 76.43,
|
| 439 |
+
"ppl": 1.50219,
|
| 440 |
+
"step": 31,
|
| 441 |
+
"tokens/total": 7466432,
|
| 442 |
+
"tokens/train_per_sec_per_gpu": 13.38,
|
| 443 |
+
"tokens/trainable": 61213
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.04251079375622717,
|
| 447 |
+
"grad_norm": 0.34962233901023865,
|
| 448 |
+
"learning_rate": 4.998130153833348e-05,
|
| 449 |
+
"loss": 0.4455469250679016,
|
| 450 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 451 |
+
"memory/max_active (GiB)": 72.26,
|
| 452 |
+
"memory/max_allocated (GiB)": 72.26,
|
| 453 |
+
"ppl": 1.56134,
|
| 454 |
+
"step": 32,
|
| 455 |
+
"tokens/total": 7682816,
|
| 456 |
+
"tokens/train_per_sec_per_gpu": 14.96,
|
| 457 |
+
"tokens/trainable": 63202
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.043839256061109264,
|
| 461 |
+
"grad_norm": 0.6062914133071899,
|
| 462 |
+
"learning_rate": 4.9976916154517596e-05,
|
| 463 |
+
"loss": 0.3847355246543884,
|
| 464 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 465 |
+
"memory/max_active (GiB)": 71.72,
|
| 466 |
+
"memory/max_allocated (GiB)": 71.72,
|
| 467 |
+
"ppl": 1.46923,
|
| 468 |
+
"step": 33,
|
| 469 |
+
"tokens/total": 7905344,
|
| 470 |
+
"tokens/train_per_sec_per_gpu": 12.46,
|
| 471 |
+
"tokens/trainable": 65086
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.045167718365991365,
|
| 475 |
+
"grad_norm": 0.25035855174064636,
|
| 476 |
+
"learning_rate": 4.997206944979284e-05,
|
| 477 |
+
"loss": 0.281414657831192,
|
| 478 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 479 |
+
"memory/max_active (GiB)": 76.48,
|
| 480 |
+
"memory/max_allocated (GiB)": 76.48,
|
| 481 |
+
"ppl": 1.325,
|
| 482 |
+
"step": 34,
|
| 483 |
+
"tokens/total": 8152960,
|
| 484 |
+
"tokens/train_per_sec_per_gpu": 10.56,
|
| 485 |
+
"tokens/trainable": 66938
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.046496180670873466,
|
| 489 |
+
"grad_norm": 0.28913062810897827,
|
| 490 |
+
"learning_rate": 4.996676151367733e-05,
|
| 491 |
+
"loss": 0.3254272937774658,
|
| 492 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 493 |
+
"memory/max_active (GiB)": 72.7,
|
| 494 |
+
"memory/max_allocated (GiB)": 72.7,
|
| 495 |
+
"ppl": 1.38462,
|
| 496 |
+
"step": 35,
|
| 497 |
+
"tokens/total": 8406912,
|
| 498 |
+
"tokens/train_per_sec_per_gpu": 11.75,
|
| 499 |
+
"tokens/trainable": 68947
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.04782464297575556,
|
| 503 |
+
"grad_norm": 0.2699942886829376,
|
| 504 |
+
"learning_rate": 4.9960992444208055e-05,
|
| 505 |
+
"loss": 0.3568383455276489,
|
| 506 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 507 |
+
"memory/max_active (GiB)": 74.66,
|
| 508 |
+
"memory/max_allocated (GiB)": 74.66,
|
| 509 |
+
"ppl": 1.4288,
|
| 510 |
+
"step": 36,
|
| 511 |
+
"tokens/total": 8657536,
|
| 512 |
+
"tokens/train_per_sec_per_gpu": 12.24,
|
| 513 |
+
"tokens/trainable": 70906
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.04915310528063766,
|
| 517 |
+
"grad_norm": 0.34783637523651123,
|
| 518 |
+
"learning_rate": 4.99547623479391e-05,
|
| 519 |
+
"loss": 0.38880521059036255,
|
| 520 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 521 |
+
"memory/max_active (GiB)": 73.35,
|
| 522 |
+
"memory/max_allocated (GiB)": 73.35,
|
| 523 |
+
"ppl": 1.47522,
|
| 524 |
+
"step": 37,
|
| 525 |
+
"tokens/total": 8891968,
|
| 526 |
+
"tokens/train_per_sec_per_gpu": 14.56,
|
| 527 |
+
"tokens/trainable": 72932
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.05048156758551976,
|
| 531 |
+
"grad_norm": 0.2874269485473633,
|
| 532 |
+
"learning_rate": 4.994807133993966e-05,
|
| 533 |
+
"loss": 0.36390942335128784,
|
| 534 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 535 |
+
"memory/max_active (GiB)": 73.19,
|
| 536 |
+
"memory/max_allocated (GiB)": 73.19,
|
| 537 |
+
"ppl": 1.43894,
|
| 538 |
+
"step": 38,
|
| 539 |
+
"tokens/total": 9137728,
|
| 540 |
+
"tokens/train_per_sec_per_gpu": 12.32,
|
| 541 |
+
"tokens/trainable": 74902
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.05181002989040186,
|
| 545 |
+
"grad_norm": 0.21942980587482452,
|
| 546 |
+
"learning_rate": 4.9940919543791934e-05,
|
| 547 |
+
"loss": 0.24741771817207336,
|
| 548 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 549 |
+
"memory/max_active (GiB)": 75.69,
|
| 550 |
+
"memory/max_allocated (GiB)": 75.69,
|
| 551 |
+
"ppl": 1.28071,
|
| 552 |
+
"step": 39,
|
| 553 |
+
"tokens/total": 9373760,
|
| 554 |
+
"tokens/train_per_sec_per_gpu": 12.65,
|
| 555 |
+
"tokens/trainable": 76811
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.05313849219528396,
|
| 559 |
+
"grad_norm": 0.2818690836429596,
|
| 560 |
+
"learning_rate": 4.9933307091588796e-05,
|
| 561 |
+
"loss": 0.3861200213432312,
|
| 562 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 563 |
+
"memory/max_active (GiB)": 73.81,
|
| 564 |
+
"memory/max_allocated (GiB)": 73.81,
|
| 565 |
+
"ppl": 1.47126,
|
| 566 |
+
"step": 40,
|
| 567 |
+
"tokens/total": 9608640,
|
| 568 |
+
"tokens/train_per_sec_per_gpu": 15.3,
|
| 569 |
+
"tokens/trainable": 79037
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.05446695450016606,
|
| 573 |
+
"grad_norm": 0.3083355724811554,
|
| 574 |
+
"learning_rate": 4.992523412393141e-05,
|
| 575 |
+
"loss": 0.3846557140350342,
|
| 576 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 577 |
+
"memory/max_active (GiB)": 73.95,
|
| 578 |
+
"memory/max_allocated (GiB)": 73.95,
|
| 579 |
+
"ppl": 1.46911,
|
| 580 |
+
"step": 41,
|
| 581 |
+
"tokens/total": 9850688,
|
| 582 |
+
"tokens/train_per_sec_per_gpu": 13.69,
|
| 583 |
+
"tokens/trainable": 81194
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.055795416805048154,
|
| 587 |
+
"grad_norm": 0.27566301822662354,
|
| 588 |
+
"learning_rate": 4.991670078992662e-05,
|
| 589 |
+
"loss": 0.375728577375412,
|
| 590 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 591 |
+
"memory/max_active (GiB)": 75.53,
|
| 592 |
+
"memory/max_allocated (GiB)": 75.53,
|
| 593 |
+
"ppl": 1.45605,
|
| 594 |
+
"step": 42,
|
| 595 |
+
"tokens/total": 10095872,
|
| 596 |
+
"tokens/train_per_sec_per_gpu": 13.84,
|
| 597 |
+
"tokens/trainable": 83205
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.057123879109930255,
|
| 601 |
+
"grad_norm": 0.2864619791507721,
|
| 602 |
+
"learning_rate": 4.990770724718415e-05,
|
| 603 |
+
"loss": 0.33605003356933594,
|
| 604 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 605 |
+
"memory/max_active (GiB)": 75.53,
|
| 606 |
+
"memory/max_allocated (GiB)": 75.53,
|
| 607 |
+
"ppl": 1.39941,
|
| 608 |
+
"step": 43,
|
| 609 |
+
"tokens/total": 10312640,
|
| 610 |
+
"tokens/train_per_sec_per_gpu": 14.12,
|
| 611 |
+
"tokens/trainable": 85157
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.058452341414812356,
|
| 615 |
+
"grad_norm": 0.33179375529289246,
|
| 616 |
+
"learning_rate": 4.9898253661813765e-05,
|
| 617 |
+
"loss": 0.386963427066803,
|
| 618 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 619 |
+
"memory/max_active (GiB)": 71.88,
|
| 620 |
+
"memory/max_allocated (GiB)": 71.88,
|
| 621 |
+
"ppl": 1.4725,
|
| 622 |
+
"step": 44,
|
| 623 |
+
"tokens/total": 10522944,
|
| 624 |
+
"tokens/train_per_sec_per_gpu": 14.61,
|
| 625 |
+
"tokens/trainable": 87062
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.05978080371969445,
|
| 629 |
+
"grad_norm": 0.2606225609779358,
|
| 630 |
+
"learning_rate": 4.9888340208422144e-05,
|
| 631 |
+
"loss": 0.2807900607585907,
|
| 632 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 633 |
+
"memory/max_active (GiB)": 73.76,
|
| 634 |
+
"memory/max_allocated (GiB)": 73.76,
|
| 635 |
+
"ppl": 1.32418,
|
| 636 |
+
"step": 45,
|
| 637 |
+
"tokens/total": 10763456,
|
| 638 |
+
"tokens/train_per_sec_per_gpu": 11.84,
|
| 639 |
+
"tokens/trainable": 88876
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.06110926602457655,
|
| 643 |
+
"grad_norm": 0.29711779952049255,
|
| 644 |
+
"learning_rate": 4.987796707010969e-05,
|
| 645 |
+
"loss": 0.3204880952835083,
|
| 646 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 647 |
+
"memory/max_active (GiB)": 76.21,
|
| 648 |
+
"memory/max_allocated (GiB)": 76.21,
|
| 649 |
+
"ppl": 1.3778,
|
| 650 |
+
"step": 46,
|
| 651 |
+
"tokens/total": 11004864,
|
| 652 |
+
"tokens/train_per_sec_per_gpu": 11.02,
|
| 653 |
+
"tokens/trainable": 90781
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.06243772832945865,
|
| 657 |
+
"grad_norm": 0.29902249574661255,
|
| 658 |
+
"learning_rate": 4.986713443846714e-05,
|
| 659 |
+
"loss": 0.34933191537857056,
|
| 660 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 661 |
+
"memory/max_active (GiB)": 76.16,
|
| 662 |
+
"memory/max_allocated (GiB)": 76.16,
|
| 663 |
+
"ppl": 1.41812,
|
| 664 |
+
"step": 47,
|
| 665 |
+
"tokens/total": 11245440,
|
| 666 |
+
"tokens/train_per_sec_per_gpu": 12.74,
|
| 667 |
+
"tokens/trainable": 92871
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.06376619063434075,
|
| 671 |
+
"grad_norm": 0.32067298889160156,
|
| 672 |
+
"learning_rate": 4.9855842513572006e-05,
|
| 673 |
+
"loss": 0.4535881280899048,
|
| 674 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 675 |
+
"memory/max_active (GiB)": 72.78,
|
| 676 |
+
"memory/max_allocated (GiB)": 72.78,
|
| 677 |
+
"ppl": 1.57395,
|
| 678 |
+
"step": 48,
|
| 679 |
+
"tokens/total": 11478272,
|
| 680 |
+
"tokens/train_per_sec_per_gpu": 14.79,
|
| 681 |
+
"tokens/trainable": 95022
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.06509465293922286,
|
| 685 |
+
"grad_norm": 0.36356621980667114,
|
| 686 |
+
"learning_rate": 4.984409150398491e-05,
|
| 687 |
+
"loss": 0.3860873579978943,
|
| 688 |
+
"memory/device_reserved (GiB)": 77.39,
|
| 689 |
+
"memory/max_active (GiB)": 73.32,
|
| 690 |
+
"memory/max_allocated (GiB)": 73.32,
|
| 691 |
+
"ppl": 1.47121,
|
| 692 |
+
"step": 49,
|
| 693 |
+
"tokens/total": 11701120,
|
| 694 |
+
"tokens/train_per_sec_per_gpu": 15.73,
|
| 695 |
+
"tokens/trainable": 97209
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.06642311524410495,
|
| 699 |
+
"grad_norm": 0.30050548911094666,
|
| 700 |
+
"learning_rate": 4.983188162674571e-05,
|
| 701 |
+
"loss": 0.4055514931678772,
|
| 702 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 703 |
+
"memory/max_active (GiB)": 76.62,
|
| 704 |
+
"memory/max_allocated (GiB)": 76.62,
|
| 705 |
+
"ppl": 1.50013,
|
| 706 |
+
"step": 50,
|
| 707 |
+
"tokens/total": 11962496,
|
| 708 |
+
"tokens/train_per_sec_per_gpu": 12.05,
|
| 709 |
+
"tokens/trainable": 99480
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.06775157754898704,
|
| 713 |
+
"grad_norm": 0.2823496162891388,
|
| 714 |
+
"learning_rate": 4.981921310736951e-05,
|
| 715 |
+
"loss": 0.3524141311645508,
|
| 716 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 717 |
+
"memory/max_active (GiB)": 71.69,
|
| 718 |
+
"memory/max_allocated (GiB)": 71.69,
|
| 719 |
+
"ppl": 1.4225,
|
| 720 |
+
"step": 51,
|
| 721 |
+
"tokens/total": 12175616,
|
| 722 |
+
"tokens/train_per_sec_per_gpu": 11.42,
|
| 723 |
+
"tokens/trainable": 101480
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.06908003985386915,
|
| 727 |
+
"grad_norm": 0.28376975655555725,
|
| 728 |
+
"learning_rate": 4.980608617984247e-05,
|
| 729 |
+
"loss": 0.34091582894325256,
|
| 730 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 731 |
+
"memory/max_active (GiB)": 76.83,
|
| 732 |
+
"memory/max_allocated (GiB)": 76.83,
|
| 733 |
+
"ppl": 1.40623,
|
| 734 |
+
"step": 52,
|
| 735 |
+
"tokens/total": 12425856,
|
| 736 |
+
"tokens/train_per_sec_per_gpu": 10.64,
|
| 737 |
+
"tokens/trainable": 103474
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.07040850215875125,
|
| 741 |
+
"grad_norm": 0.3835979998111725,
|
| 742 |
+
"learning_rate": 4.97925010866175e-05,
|
| 743 |
+
"loss": 0.28788986802101135,
|
| 744 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 745 |
+
"memory/max_active (GiB)": 74.28,
|
| 746 |
+
"memory/max_allocated (GiB)": 74.28,
|
| 747 |
+
"ppl": 1.33361,
|
| 748 |
+
"step": 53,
|
| 749 |
+
"tokens/total": 12673472,
|
| 750 |
+
"tokens/train_per_sec_per_gpu": 10.97,
|
| 751 |
+
"tokens/trainable": 105399
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.07173696446363334,
|
| 755 |
+
"grad_norm": 0.39083245396614075,
|
| 756 |
+
"learning_rate": 4.97784580786098e-05,
|
| 757 |
+
"loss": 0.3275277316570282,
|
| 758 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 759 |
+
"memory/max_active (GiB)": 75.58,
|
| 760 |
+
"memory/max_allocated (GiB)": 75.58,
|
| 761 |
+
"ppl": 1.38753,
|
| 762 |
+
"step": 54,
|
| 763 |
+
"tokens/total": 12919936,
|
| 764 |
+
"tokens/train_per_sec_per_gpu": 12.12,
|
| 765 |
+
"tokens/trainable": 107320
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.07306542676851545,
|
| 769 |
+
"grad_norm": 0.28509923815727234,
|
| 770 |
+
"learning_rate": 4.9763957415192175e-05,
|
| 771 |
+
"loss": 0.288000226020813,
|
| 772 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 773 |
+
"memory/max_active (GiB)": 71.94,
|
| 774 |
+
"memory/max_allocated (GiB)": 71.94,
|
| 775 |
+
"ppl": 1.33376,
|
| 776 |
+
"step": 55,
|
| 777 |
+
"tokens/total": 13149568,
|
| 778 |
+
"tokens/train_per_sec_per_gpu": 12.46,
|
| 779 |
+
"tokens/trainable": 109265
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.07439388907339754,
|
| 783 |
+
"grad_norm": 0.2659589946269989,
|
| 784 |
+
"learning_rate": 4.9748999364190304e-05,
|
| 785 |
+
"loss": 0.3212144374847412,
|
| 786 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 787 |
+
"memory/max_active (GiB)": 75.94,
|
| 788 |
+
"memory/max_allocated (GiB)": 75.94,
|
| 789 |
+
"ppl": 1.3788,
|
| 790 |
+
"step": 56,
|
| 791 |
+
"tokens/total": 13412032,
|
| 792 |
+
"tokens/train_per_sec_per_gpu": 11.09,
|
| 793 |
+
"tokens/trainable": 111171
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.07572235137827964,
|
| 797 |
+
"grad_norm": 0.2954414486885071,
|
| 798 |
+
"learning_rate": 4.973358420187776e-05,
|
| 799 |
+
"loss": 0.35858094692230225,
|
| 800 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 801 |
+
"memory/max_active (GiB)": 74.44,
|
| 802 |
+
"memory/max_allocated (GiB)": 74.44,
|
| 803 |
+
"ppl": 1.4313,
|
| 804 |
+
"step": 57,
|
| 805 |
+
"tokens/total": 13670784,
|
| 806 |
+
"tokens/train_per_sec_per_gpu": 13.99,
|
| 807 |
+
"tokens/trainable": 113207
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.07705081368316175,
|
| 811 |
+
"grad_norm": 0.29742714762687683,
|
| 812 |
+
"learning_rate": 4.971771221297088e-05,
|
| 813 |
+
"loss": 0.3568863570690155,
|
| 814 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 815 |
+
"memory/max_active (GiB)": 74.6,
|
| 816 |
+
"memory/max_allocated (GiB)": 74.6,
|
| 817 |
+
"ppl": 1.42887,
|
| 818 |
+
"step": 58,
|
| 819 |
+
"tokens/total": 13901888,
|
| 820 |
+
"tokens/train_per_sec_per_gpu": 13.24,
|
| 821 |
+
"tokens/trainable": 115221
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.07837927598804384,
|
| 825 |
+
"grad_norm": 0.5640520453453064,
|
| 826 |
+
"learning_rate": 4.970138369062359e-05,
|
| 827 |
+
"loss": 0.28157147765159607,
|
| 828 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 829 |
+
"memory/max_active (GiB)": 76.02,
|
| 830 |
+
"memory/max_allocated (GiB)": 76.02,
|
| 831 |
+
"ppl": 1.32521,
|
| 832 |
+
"step": 59,
|
| 833 |
+
"tokens/total": 14140224,
|
| 834 |
+
"tokens/train_per_sec_per_gpu": 11.3,
|
| 835 |
+
"tokens/trainable": 117086
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.07970773829292593,
|
| 839 |
+
"grad_norm": 0.28851282596588135,
|
| 840 |
+
"learning_rate": 4.968459893642189e-05,
|
| 841 |
+
"loss": 0.34128955006599426,
|
| 842 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 843 |
+
"memory/max_active (GiB)": 74.6,
|
| 844 |
+
"memory/max_allocated (GiB)": 74.6,
|
| 845 |
+
"ppl": 1.40676,
|
| 846 |
+
"step": 60,
|
| 847 |
+
"tokens/total": 14384256,
|
| 848 |
+
"tokens/train_per_sec_per_gpu": 14.41,
|
| 849 |
+
"tokens/trainable": 119322
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.08103620059780804,
|
| 853 |
+
"grad_norm": 0.3004905879497528,
|
| 854 |
+
"learning_rate": 4.966735826037836e-05,
|
| 855 |
+
"loss": 0.34793412685394287,
|
| 856 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 857 |
+
"memory/max_active (GiB)": 75.72,
|
| 858 |
+
"memory/max_allocated (GiB)": 75.72,
|
| 859 |
+
"ppl": 1.41614,
|
| 860 |
+
"step": 61,
|
| 861 |
+
"tokens/total": 14647168,
|
| 862 |
+
"tokens/train_per_sec_per_gpu": 11.17,
|
| 863 |
+
"tokens/trainable": 121273
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.08236466290269014,
|
| 867 |
+
"grad_norm": 0.656767725944519,
|
| 868 |
+
"learning_rate": 4.9649661980926406e-05,
|
| 869 |
+
"loss": 0.3032311797142029,
|
| 870 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 871 |
+
"memory/max_active (GiB)": 75.26,
|
| 872 |
+
"memory/max_allocated (GiB)": 75.26,
|
| 873 |
+
"ppl": 1.35423,
|
| 874 |
+
"step": 62,
|
| 875 |
+
"tokens/total": 14905792,
|
| 876 |
+
"tokens/train_per_sec_per_gpu": 13.11,
|
| 877 |
+
"tokens/trainable": 123323
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.08369312520757223,
|
| 881 |
+
"grad_norm": 0.3169984221458435,
|
| 882 |
+
"learning_rate": 4.963151042491437e-05,
|
| 883 |
+
"loss": 0.365967720746994,
|
| 884 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 885 |
+
"memory/max_active (GiB)": 75.28,
|
| 886 |
+
"memory/max_allocated (GiB)": 75.28,
|
| 887 |
+
"ppl": 1.44191,
|
| 888 |
+
"step": 63,
|
| 889 |
+
"tokens/total": 15153664,
|
| 890 |
+
"tokens/train_per_sec_per_gpu": 10.93,
|
| 891 |
+
"tokens/trainable": 125382
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.08502158751245434,
|
| 895 |
+
"grad_norm": 0.33295902609825134,
|
| 896 |
+
"learning_rate": 4.9612903927599505e-05,
|
| 897 |
+
"loss": 0.3628847599029541,
|
| 898 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 899 |
+
"memory/max_active (GiB)": 72.29,
|
| 900 |
+
"memory/max_allocated (GiB)": 72.29,
|
| 901 |
+
"ppl": 1.43747,
|
| 902 |
+
"step": 64,
|
| 903 |
+
"tokens/total": 15367552,
|
| 904 |
+
"tokens/train_per_sec_per_gpu": 15.24,
|
| 905 |
+
"tokens/trainable": 127310
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.08635004981733643,
|
| 909 |
+
"grad_norm": 0.41118544340133667,
|
| 910 |
+
"learning_rate": 4.959384283264179e-05,
|
| 911 |
+
"loss": 0.41343849897384644,
|
| 912 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 913 |
+
"memory/max_active (GiB)": 72.56,
|
| 914 |
+
"memory/max_allocated (GiB)": 72.56,
|
| 915 |
+
"ppl": 1.51201,
|
| 916 |
+
"step": 65,
|
| 917 |
+
"tokens/total": 15590848,
|
| 918 |
+
"tokens/train_per_sec_per_gpu": 14.13,
|
| 919 |
+
"tokens/trainable": 129279
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.08767851212221853,
|
| 923 |
+
"grad_norm": 0.29445692896842957,
|
| 924 |
+
"learning_rate": 4.957432749209755e-05,
|
| 925 |
+
"loss": 0.34112682938575745,
|
| 926 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 927 |
+
"memory/max_active (GiB)": 73.44,
|
| 928 |
+
"memory/max_allocated (GiB)": 73.44,
|
| 929 |
+
"ppl": 1.40653,
|
| 930 |
+
"step": 66,
|
| 931 |
+
"tokens/total": 15858432,
|
| 932 |
+
"tokens/train_per_sec_per_gpu": 11.05,
|
| 933 |
+
"tokens/trainable": 131138
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.08900697442710064,
|
| 937 |
+
"grad_norm": 0.3269948661327362,
|
| 938 |
+
"learning_rate": 4.955435826641299e-05,
|
| 939 |
+
"loss": 0.3915187120437622,
|
| 940 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 941 |
+
"memory/max_active (GiB)": 74.88,
|
| 942 |
+
"memory/max_allocated (GiB)": 74.88,
|
| 943 |
+
"ppl": 1.47923,
|
| 944 |
+
"step": 67,
|
| 945 |
+
"tokens/total": 16085504,
|
| 946 |
+
"tokens/train_per_sec_per_gpu": 17.46,
|
| 947 |
+
"tokens/trainable": 133359
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.09033543673198273,
|
| 951 |
+
"grad_norm": 0.30889901518821716,
|
| 952 |
+
"learning_rate": 4.953393552441753e-05,
|
| 953 |
+
"loss": 0.2897987961769104,
|
| 954 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 955 |
+
"memory/max_active (GiB)": 74.74,
|
| 956 |
+
"memory/max_allocated (GiB)": 74.74,
|
| 957 |
+
"ppl": 1.33616,
|
| 958 |
+
"step": 68,
|
| 959 |
+
"tokens/total": 16326272,
|
| 960 |
+
"tokens/train_per_sec_per_gpu": 12.32,
|
| 961 |
+
"tokens/trainable": 135250
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.09166389903686482,
|
| 965 |
+
"grad_norm": 0.3052501082420349,
|
| 966 |
+
"learning_rate": 4.951305964331698e-05,
|
| 967 |
+
"loss": 0.2914356589317322,
|
| 968 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 969 |
+
"memory/max_active (GiB)": 75.96,
|
| 970 |
+
"memory/max_allocated (GiB)": 75.96,
|
| 971 |
+
"ppl": 1.33835,
|
| 972 |
+
"step": 69,
|
| 973 |
+
"tokens/total": 16568384,
|
| 974 |
+
"tokens/train_per_sec_per_gpu": 12.94,
|
| 975 |
+
"tokens/trainable": 137383
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.09299236134174693,
|
| 979 |
+
"grad_norm": 0.2697888910770416,
|
| 980 |
+
"learning_rate": 4.9491731008686584e-05,
|
| 981 |
+
"loss": 0.2577957212924957,
|
| 982 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 983 |
+
"memory/max_active (GiB)": 75.47,
|
| 984 |
+
"memory/max_allocated (GiB)": 75.47,
|
| 985 |
+
"ppl": 1.29407,
|
| 986 |
+
"step": 70,
|
| 987 |
+
"tokens/total": 16793536,
|
| 988 |
+
"tokens/train_per_sec_per_gpu": 13.0,
|
| 989 |
+
"tokens/trainable": 139172
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.09432082364662903,
|
| 993 |
+
"grad_norm": 0.29883286356925964,
|
| 994 |
+
"learning_rate": 4.9469950014463876e-05,
|
| 995 |
+
"loss": 0.30956798791885376,
|
| 996 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 997 |
+
"memory/max_active (GiB)": 74.82,
|
| 998 |
+
"memory/max_allocated (GiB)": 74.82,
|
| 999 |
+
"ppl": 1.36284,
|
| 1000 |
+
"step": 71,
|
| 1001 |
+
"tokens/total": 17018816,
|
| 1002 |
+
"tokens/train_per_sec_per_gpu": 13.42,
|
| 1003 |
+
"tokens/trainable": 141287
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.09564928595151112,
|
| 1007 |
+
"grad_norm": 0.26738041639328003,
|
| 1008 |
+
"learning_rate": 4.944771706294145e-05,
|
| 1009 |
+
"loss": 0.306536465883255,
|
| 1010 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 1011 |
+
"memory/max_active (GiB)": 74.39,
|
| 1012 |
+
"memory/max_allocated (GiB)": 74.39,
|
| 1013 |
+
"ppl": 1.35871,
|
| 1014 |
+
"step": 72,
|
| 1015 |
+
"tokens/total": 17269248,
|
| 1016 |
+
"tokens/train_per_sec_per_gpu": 11.76,
|
| 1017 |
+
"tokens/trainable": 143236
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.09697774825639323,
|
| 1021 |
+
"grad_norm": 0.3274018168449402,
|
| 1022 |
+
"learning_rate": 4.9425032564759485e-05,
|
| 1023 |
+
"loss": 0.34317129850387573,
|
| 1024 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 1025 |
+
"memory/max_active (GiB)": 76.24,
|
| 1026 |
+
"memory/max_allocated (GiB)": 76.24,
|
| 1027 |
+
"ppl": 1.40941,
|
| 1028 |
+
"step": 73,
|
| 1029 |
+
"tokens/total": 17508992,
|
| 1030 |
+
"tokens/train_per_sec_per_gpu": 14.64,
|
| 1031 |
+
"tokens/trainable": 145571
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.09830621056127532,
|
| 1035 |
+
"grad_norm": 0.27942293882369995,
|
| 1036 |
+
"learning_rate": 4.9401896938898185e-05,
|
| 1037 |
+
"loss": 0.32496753334999084,
|
| 1038 |
+
"memory/device_reserved (GiB)": 77.59,
|
| 1039 |
+
"memory/max_active (GiB)": 74.9,
|
| 1040 |
+
"memory/max_allocated (GiB)": 74.9,
|
| 1041 |
+
"ppl": 1.38399,
|
| 1042 |
+
"step": 74,
|
| 1043 |
+
"tokens/total": 17736448,
|
| 1044 |
+
"tokens/train_per_sec_per_gpu": 13.25,
|
| 1045 |
+
"tokens/trainable": 147614
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.09963467286615742,
|
| 1049 |
+
"grad_norm": 0.20825843513011932,
|
| 1050 |
+
"learning_rate": 4.937831061267001e-05,
|
| 1051 |
+
"loss": 0.22000554203987122,
|
| 1052 |
+
"memory/device_reserved (GiB)": 77.79,
|
| 1053 |
+
"memory/max_active (GiB)": 76.86,
|
| 1054 |
+
"memory/max_allocated (GiB)": 76.86,
|
| 1055 |
+
"ppl": 1.24608,
|
| 1056 |
+
"step": 75,
|
| 1057 |
+
"tokens/total": 18018752,
|
| 1058 |
+
"tokens/train_per_sec_per_gpu": 11.4,
|
| 1059 |
+
"tokens/trainable": 149818
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.10096313517103953,
|
| 1063 |
+
"grad_norm": 0.28913629055023193,
|
| 1064 |
+
"learning_rate": 4.935427402171185e-05,
|
| 1065 |
+
"loss": 0.34746575355529785,
|
| 1066 |
+
"memory/device_reserved (GiB)": 77.79,
|
| 1067 |
+
"memory/max_active (GiB)": 75.58,
|
| 1068 |
+
"memory/max_allocated (GiB)": 75.58,
|
| 1069 |
+
"ppl": 1.41548,
|
| 1070 |
+
"step": 76,
|
| 1071 |
+
"tokens/total": 18233856,
|
| 1072 |
+
"tokens/train_per_sec_per_gpu": 9.96,
|
| 1073 |
+
"tokens/trainable": 151707
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.10229159747592162,
|
| 1077 |
+
"grad_norm": 0.25012022256851196,
|
| 1078 |
+
"learning_rate": 4.9329787609976893e-05,
|
| 1079 |
+
"loss": 0.28436708450317383,
|
| 1080 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1081 |
+
"memory/max_active (GiB)": 76.32,
|
| 1082 |
+
"memory/max_allocated (GiB)": 76.32,
|
| 1083 |
+
"ppl": 1.32892,
|
| 1084 |
+
"step": 77,
|
| 1085 |
+
"tokens/total": 18452672,
|
| 1086 |
+
"tokens/train_per_sec_per_gpu": 12.01,
|
| 1087 |
+
"tokens/trainable": 153598
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.10362005978080371,
|
| 1091 |
+
"grad_norm": 0.49168774485588074,
|
| 1092 |
+
"learning_rate": 4.9304851829726494e-05,
|
| 1093 |
+
"loss": 0.28677767515182495,
|
| 1094 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1095 |
+
"memory/max_active (GiB)": 75.34,
|
| 1096 |
+
"memory/max_allocated (GiB)": 75.34,
|
| 1097 |
+
"ppl": 1.33213,
|
| 1098 |
+
"step": 78,
|
| 1099 |
+
"tokens/total": 18695552,
|
| 1100 |
+
"tokens/train_per_sec_per_gpu": 12.31,
|
| 1101 |
+
"tokens/trainable": 155561
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.10494852208568582,
|
| 1105 |
+
"grad_norm": 0.5867858529090881,
|
| 1106 |
+
"learning_rate": 4.927946714152178e-05,
|
| 1107 |
+
"loss": 0.38911110162734985,
|
| 1108 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1109 |
+
"memory/max_active (GiB)": 74.52,
|
| 1110 |
+
"memory/max_allocated (GiB)": 74.52,
|
| 1111 |
+
"ppl": 1.47567,
|
| 1112 |
+
"step": 79,
|
| 1113 |
+
"tokens/total": 18907712,
|
| 1114 |
+
"tokens/train_per_sec_per_gpu": 15.22,
|
| 1115 |
+
"tokens/trainable": 157505
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.10627698439056792,
|
| 1119 |
+
"grad_norm": 0.27808260917663574,
|
| 1120 |
+
"learning_rate": 4.925363401421517e-05,
|
| 1121 |
+
"loss": 0.3041893541812897,
|
| 1122 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1123 |
+
"memory/max_active (GiB)": 76.4,
|
| 1124 |
+
"memory/max_allocated (GiB)": 76.4,
|
| 1125 |
+
"ppl": 1.35553,
|
| 1126 |
+
"step": 80,
|
| 1127 |
+
"tokens/total": 19134144,
|
| 1128 |
+
"tokens/train_per_sec_per_gpu": 12.57,
|
| 1129 |
+
"tokens/trainable": 159379
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.10760544669545001,
|
| 1133 |
+
"grad_norm": 0.2894990146160126,
|
| 1134 |
+
"learning_rate": 4.92273529249417e-05,
|
| 1135 |
+
"loss": 0.3562927842140198,
|
| 1136 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1137 |
+
"memory/max_active (GiB)": 72.18,
|
| 1138 |
+
"memory/max_allocated (GiB)": 72.18,
|
| 1139 |
+
"ppl": 1.42803,
|
| 1140 |
+
"step": 81,
|
| 1141 |
+
"tokens/total": 19376000,
|
| 1142 |
+
"tokens/train_per_sec_per_gpu": 12.75,
|
| 1143 |
+
"tokens/trainable": 161333
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.10893390900033212,
|
| 1147 |
+
"grad_norm": 0.28392910957336426,
|
| 1148 |
+
"learning_rate": 4.9200624359110206e-05,
|
| 1149 |
+
"loss": 0.32886797189712524,
|
| 1150 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1151 |
+
"memory/max_active (GiB)": 74.09,
|
| 1152 |
+
"memory/max_allocated (GiB)": 74.09,
|
| 1153 |
+
"ppl": 1.38939,
|
| 1154 |
+
"step": 82,
|
| 1155 |
+
"tokens/total": 19613056,
|
| 1156 |
+
"tokens/train_per_sec_per_gpu": 16.55,
|
| 1157 |
+
"tokens/trainable": 163537
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.11026237130521421,
|
| 1161 |
+
"grad_norm": 0.272851824760437,
|
| 1162 |
+
"learning_rate": 4.9173448810394385e-05,
|
| 1163 |
+
"loss": 0.25960707664489746,
|
| 1164 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1165 |
+
"memory/max_active (GiB)": 75.37,
|
| 1166 |
+
"memory/max_allocated (GiB)": 75.37,
|
| 1167 |
+
"ppl": 1.29642,
|
| 1168 |
+
"step": 83,
|
| 1169 |
+
"tokens/total": 19840960,
|
| 1170 |
+
"tokens/train_per_sec_per_gpu": 10.34,
|
| 1171 |
+
"tokens/trainable": 165415
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.11159083361009631,
|
| 1175 |
+
"grad_norm": 0.30733823776245117,
|
| 1176 |
+
"learning_rate": 4.9145826780723634e-05,
|
| 1177 |
+
"loss": 0.29021644592285156,
|
| 1178 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1179 |
+
"memory/max_active (GiB)": 74.69,
|
| 1180 |
+
"memory/max_allocated (GiB)": 74.69,
|
| 1181 |
+
"ppl": 1.33672,
|
| 1182 |
+
"step": 84,
|
| 1183 |
+
"tokens/total": 20084608,
|
| 1184 |
+
"tokens/train_per_sec_per_gpu": 10.81,
|
| 1185 |
+
"tokens/trainable": 167156
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.11291929591497842,
|
| 1189 |
+
"grad_norm": 0.24903596937656403,
|
| 1190 |
+
"learning_rate": 4.911775878027384e-05,
|
| 1191 |
+
"loss": 0.25815725326538086,
|
| 1192 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1193 |
+
"memory/max_active (GiB)": 73.24,
|
| 1194 |
+
"memory/max_allocated (GiB)": 73.24,
|
| 1195 |
+
"ppl": 1.29454,
|
| 1196 |
+
"step": 85,
|
| 1197 |
+
"tokens/total": 20300224,
|
| 1198 |
+
"tokens/train_per_sec_per_gpu": 13.75,
|
| 1199 |
+
"tokens/trainable": 169157
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.11424775821986051,
|
| 1203 |
+
"grad_norm": 0.2804400324821472,
|
| 1204 |
+
"learning_rate": 4.9089245327457876e-05,
|
| 1205 |
+
"loss": 0.3358738124370575,
|
| 1206 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1207 |
+
"memory/max_active (GiB)": 73.6,
|
| 1208 |
+
"memory/max_allocated (GiB)": 73.6,
|
| 1209 |
+
"ppl": 1.39916,
|
| 1210 |
+
"step": 86,
|
| 1211 |
+
"tokens/total": 20541056,
|
| 1212 |
+
"tokens/train_per_sec_per_gpu": 13.75,
|
| 1213 |
+
"tokens/trainable": 171164
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.1155762205247426,
|
| 1217 |
+
"grad_norm": 0.3141542375087738,
|
| 1218 |
+
"learning_rate": 4.906028694891612e-05,
|
| 1219 |
+
"loss": 0.3537227511405945,
|
| 1220 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1221 |
+
"memory/max_active (GiB)": 76.13,
|
| 1222 |
+
"memory/max_allocated (GiB)": 76.13,
|
| 1223 |
+
"ppl": 1.42436,
|
| 1224 |
+
"step": 87,
|
| 1225 |
+
"tokens/total": 20785216,
|
| 1226 |
+
"tokens/train_per_sec_per_gpu": 13.88,
|
| 1227 |
+
"tokens/trainable": 173239
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.11690468282962471,
|
| 1231 |
+
"grad_norm": 0.2685106098651886,
|
| 1232 |
+
"learning_rate": 4.903088417950664e-05,
|
| 1233 |
+
"loss": 0.27877840399742126,
|
| 1234 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1235 |
+
"memory/max_active (GiB)": 72.67,
|
| 1236 |
+
"memory/max_allocated (GiB)": 72.67,
|
| 1237 |
+
"ppl": 1.32151,
|
| 1238 |
+
"step": 88,
|
| 1239 |
+
"tokens/total": 21016768,
|
| 1240 |
+
"tokens/train_per_sec_per_gpu": 12.39,
|
| 1241 |
+
"tokens/trainable": 175096
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.1182331451345068,
|
| 1245 |
+
"grad_norm": 0.3396528661251068,
|
| 1246 |
+
"learning_rate": 4.900103756229538e-05,
|
| 1247 |
+
"loss": 0.31499725580215454,
|
| 1248 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1249 |
+
"memory/max_active (GiB)": 74.71,
|
| 1250 |
+
"memory/max_allocated (GiB)": 74.71,
|
| 1251 |
+
"ppl": 1.37026,
|
| 1252 |
+
"step": 89,
|
| 1253 |
+
"tokens/total": 21246976,
|
| 1254 |
+
"tokens/train_per_sec_per_gpu": 13.24,
|
| 1255 |
+
"tokens/trainable": 177111
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.1195616074393889,
|
| 1259 |
+
"grad_norm": 0.2889237403869629,
|
| 1260 |
+
"learning_rate": 4.897074764854608e-05,
|
| 1261 |
+
"loss": 0.3066895306110382,
|
| 1262 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1263 |
+
"memory/max_active (GiB)": 74.09,
|
| 1264 |
+
"memory/max_allocated (GiB)": 74.09,
|
| 1265 |
+
"ppl": 1.35892,
|
| 1266 |
+
"step": 90,
|
| 1267 |
+
"tokens/total": 21491072,
|
| 1268 |
+
"tokens/train_per_sec_per_gpu": 12.02,
|
| 1269 |
+
"tokens/trainable": 179077
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.12089006974427101,
|
| 1273 |
+
"grad_norm": 0.2633538842201233,
|
| 1274 |
+
"learning_rate": 4.894001499771015e-05,
|
| 1275 |
+
"loss": 0.3202817440032959,
|
| 1276 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1277 |
+
"memory/max_active (GiB)": 75.86,
|
| 1278 |
+
"memory/max_allocated (GiB)": 75.86,
|
| 1279 |
+
"ppl": 1.37752,
|
| 1280 |
+
"step": 91,
|
| 1281 |
+
"tokens/total": 21734208,
|
| 1282 |
+
"tokens/train_per_sec_per_gpu": 13.09,
|
| 1283 |
+
"tokens/trainable": 181123
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.1222185320491531,
|
| 1287 |
+
"grad_norm": 0.28408122062683105,
|
| 1288 |
+
"learning_rate": 4.890884017741628e-05,
|
| 1289 |
+
"loss": 0.30145886540412903,
|
| 1290 |
+
"memory/device_reserved (GiB)": 77.2,
|
| 1291 |
+
"memory/max_active (GiB)": 75.53,
|
| 1292 |
+
"memory/max_allocated (GiB)": 75.53,
|
| 1293 |
+
"ppl": 1.35183,
|
| 1294 |
+
"step": 92,
|
| 1295 |
+
"tokens/total": 21955776,
|
| 1296 |
+
"tokens/train_per_sec_per_gpu": 15.96,
|
| 1297 |
+
"tokens/trainable": 183236
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.12354699435403521,
|
| 1301 |
+
"grad_norm": 0.2829677164554596,
|
| 1302 |
+
"learning_rate": 4.887722376345999e-05,
|
| 1303 |
+
"loss": 0.386635959148407,
|
| 1304 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1305 |
+
"memory/max_active (GiB)": 76.84,
|
| 1306 |
+
"memory/max_allocated (GiB)": 76.84,
|
| 1307 |
+
"ppl": 1.47202,
|
| 1308 |
+
"step": 93,
|
| 1309 |
+
"tokens/total": 22220800,
|
| 1310 |
+
"tokens/train_per_sec_per_gpu": 13.65,
|
| 1311 |
+
"tokens/trainable": 185470
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.1248754566589173,
|
| 1315 |
+
"grad_norm": 0.2762235701084137,
|
| 1316 |
+
"learning_rate": 4.884516633979299e-05,
|
| 1317 |
+
"loss": 0.2841187119483948,
|
| 1318 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1319 |
+
"memory/max_active (GiB)": 76.16,
|
| 1320 |
+
"memory/max_allocated (GiB)": 76.16,
|
| 1321 |
+
"ppl": 1.32859,
|
| 1322 |
+
"step": 94,
|
| 1323 |
+
"tokens/total": 22472000,
|
| 1324 |
+
"tokens/train_per_sec_per_gpu": 12.02,
|
| 1325 |
+
"tokens/trainable": 187632
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.1262039189637994,
|
| 1329 |
+
"grad_norm": 0.27298712730407715,
|
| 1330 |
+
"learning_rate": 4.881266849851239e-05,
|
| 1331 |
+
"loss": 0.31646984815597534,
|
| 1332 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1333 |
+
"memory/max_active (GiB)": 72.1,
|
| 1334 |
+
"memory/max_allocated (GiB)": 72.1,
|
| 1335 |
+
"ppl": 1.37227,
|
| 1336 |
+
"step": 95,
|
| 1337 |
+
"tokens/total": 22724928,
|
| 1338 |
+
"tokens/train_per_sec_per_gpu": 10.78,
|
| 1339 |
+
"tokens/trainable": 189511
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.1275323812686815,
|
| 1343 |
+
"grad_norm": 0.30283185839653015,
|
| 1344 |
+
"learning_rate": 4.877973083984976e-05,
|
| 1345 |
+
"loss": 0.32183483242988586,
|
| 1346 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1347 |
+
"memory/max_active (GiB)": 73.43,
|
| 1348 |
+
"memory/max_allocated (GiB)": 73.43,
|
| 1349 |
+
"ppl": 1.37966,
|
| 1350 |
+
"step": 96,
|
| 1351 |
+
"tokens/total": 22930688,
|
| 1352 |
+
"tokens/train_per_sec_per_gpu": 17.58,
|
| 1353 |
+
"tokens/trainable": 191527
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.1288608435735636,
|
| 1357 |
+
"grad_norm": 0.39260604977607727,
|
| 1358 |
+
"learning_rate": 4.874635397216005e-05,
|
| 1359 |
+
"loss": 0.30737465620040894,
|
| 1360 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1361 |
+
"memory/max_active (GiB)": 72.92,
|
| 1362 |
+
"memory/max_allocated (GiB)": 72.92,
|
| 1363 |
+
"ppl": 1.35985,
|
| 1364 |
+
"step": 97,
|
| 1365 |
+
"tokens/total": 23159808,
|
| 1366 |
+
"tokens/train_per_sec_per_gpu": 13.92,
|
| 1367 |
+
"tokens/trainable": 193452
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.1301893058784457,
|
| 1371 |
+
"grad_norm": 0.3084384799003601,
|
| 1372 |
+
"learning_rate": 4.8712538511910354e-05,
|
| 1373 |
+
"loss": 0.2816072404384613,
|
| 1374 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1375 |
+
"memory/max_active (GiB)": 76.51,
|
| 1376 |
+
"memory/max_allocated (GiB)": 76.51,
|
| 1377 |
+
"ppl": 1.32526,
|
| 1378 |
+
"step": 98,
|
| 1379 |
+
"tokens/total": 23417536,
|
| 1380 |
+
"tokens/train_per_sec_per_gpu": 11.32,
|
| 1381 |
+
"tokens/trainable": 195481
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.1315177681833278,
|
| 1385 |
+
"grad_norm": 0.3135785460472107,
|
| 1386 |
+
"learning_rate": 4.867828508366853e-05,
|
| 1387 |
+
"loss": 0.3494204878807068,
|
| 1388 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1389 |
+
"memory/max_active (GiB)": 73.65,
|
| 1390 |
+
"memory/max_allocated (GiB)": 73.65,
|
| 1391 |
+
"ppl": 1.41825,
|
| 1392 |
+
"step": 99,
|
| 1393 |
+
"tokens/total": 23646592,
|
| 1394 |
+
"tokens/train_per_sec_per_gpu": 15.39,
|
| 1395 |
+
"tokens/trainable": 197676
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.1328462304882099,
|
| 1399 |
+
"grad_norm": 0.40487584471702576,
|
| 1400 |
+
"learning_rate": 4.864359432009164e-05,
|
| 1401 |
+
"loss": 0.36197125911712646,
|
| 1402 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1403 |
+
"memory/max_active (GiB)": 76.02,
|
| 1404 |
+
"memory/max_allocated (GiB)": 76.02,
|
| 1405 |
+
"ppl": 1.43616,
|
| 1406 |
+
"step": 100,
|
| 1407 |
+
"tokens/total": 23892672,
|
| 1408 |
+
"tokens/train_per_sec_per_gpu": 11.81,
|
| 1409 |
+
"tokens/trainable": 199656
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.134174692793092,
|
| 1413 |
+
"grad_norm": 0.31580036878585815,
|
| 1414 |
+
"learning_rate": 4.860846686191431e-05,
|
| 1415 |
+
"loss": 0.31181108951568604,
|
| 1416 |
+
"memory/device_reserved (GiB)": 77.55,
|
| 1417 |
+
"memory/max_active (GiB)": 74.41,
|
| 1418 |
+
"memory/max_allocated (GiB)": 74.41,
|
| 1419 |
+
"ppl": 1.3659,
|
| 1420 |
+
"step": 101,
|
| 1421 |
+
"tokens/total": 24151808,
|
| 1422 |
+
"tokens/train_per_sec_per_gpu": 9.4,
|
| 1423 |
+
"tokens/trainable": 201706
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.1355031550979741,
|
| 1427 |
+
"grad_norm": 0.32332488894462585,
|
| 1428 |
+
"learning_rate": 4.857290335793683e-05,
|
| 1429 |
+
"loss": 0.3490774631500244,
|
| 1430 |
+
"memory/device_reserved (GiB)": 76.51,
|
| 1431 |
+
"memory/max_active (GiB)": 75.5,
|
| 1432 |
+
"memory/max_allocated (GiB)": 75.5,
|
| 1433 |
+
"ppl": 1.41776,
|
| 1434 |
+
"step": 102,
|
| 1435 |
+
"tokens/total": 24394816,
|
| 1436 |
+
"tokens/train_per_sec_per_gpu": 13.6,
|
| 1437 |
+
"tokens/trainable": 203823
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.13683161740285618,
|
| 1441 |
+
"grad_norm": 0.3062637746334076,
|
| 1442 |
+
"learning_rate": 4.853690446501323e-05,
|
| 1443 |
+
"loss": 0.34347471594810486,
|
| 1444 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1445 |
+
"memory/max_active (GiB)": 76.86,
|
| 1446 |
+
"memory/max_allocated (GiB)": 76.86,
|
| 1447 |
+
"ppl": 1.40984,
|
| 1448 |
+
"step": 103,
|
| 1449 |
+
"tokens/total": 24640512,
|
| 1450 |
+
"tokens/train_per_sec_per_gpu": 12.27,
|
| 1451 |
+
"tokens/trainable": 205892
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.1381600797077383,
|
| 1455 |
+
"grad_norm": 0.31956177949905396,
|
| 1456 |
+
"learning_rate": 4.850047084803914e-05,
|
| 1457 |
+
"loss": 0.3540318012237549,
|
| 1458 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1459 |
+
"memory/max_active (GiB)": 74.41,
|
| 1460 |
+
"memory/max_allocated (GiB)": 74.41,
|
| 1461 |
+
"ppl": 1.4248,
|
| 1462 |
+
"step": 104,
|
| 1463 |
+
"tokens/total": 24887680,
|
| 1464 |
+
"tokens/train_per_sec_per_gpu": 14.75,
|
| 1465 |
+
"tokens/trainable": 208306
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.1394885420126204,
|
| 1469 |
+
"grad_norm": 0.2798771560192108,
|
| 1470 |
+
"learning_rate": 4.846360317993947e-05,
|
| 1471 |
+
"loss": 0.3245096206665039,
|
| 1472 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1473 |
+
"memory/max_active (GiB)": 75.31,
|
| 1474 |
+
"memory/max_allocated (GiB)": 75.31,
|
| 1475 |
+
"ppl": 1.38335,
|
| 1476 |
+
"step": 105,
|
| 1477 |
+
"tokens/total": 25134848,
|
| 1478 |
+
"tokens/train_per_sec_per_gpu": 12.61,
|
| 1479 |
+
"tokens/trainable": 210329
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.1408170043175025,
|
| 1483 |
+
"grad_norm": 0.3507809042930603,
|
| 1484 |
+
"learning_rate": 4.842630214165601e-05,
|
| 1485 |
+
"loss": 0.38814184069633484,
|
| 1486 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1487 |
+
"memory/max_active (GiB)": 72.4,
|
| 1488 |
+
"memory/max_allocated (GiB)": 72.4,
|
| 1489 |
+
"ppl": 1.47424,
|
| 1490 |
+
"step": 106,
|
| 1491 |
+
"tokens/total": 25371904,
|
| 1492 |
+
"tokens/train_per_sec_per_gpu": 14.97,
|
| 1493 |
+
"tokens/trainable": 212511
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.1421454666223846,
|
| 1497 |
+
"grad_norm": 0.26299262046813965,
|
| 1498 |
+
"learning_rate": 4.838856842213486e-05,
|
| 1499 |
+
"loss": 0.2627770006656647,
|
| 1500 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1501 |
+
"memory/max_active (GiB)": 73.71,
|
| 1502 |
+
"memory/max_allocated (GiB)": 73.71,
|
| 1503 |
+
"ppl": 1.30054,
|
| 1504 |
+
"step": 107,
|
| 1505 |
+
"tokens/total": 25620672,
|
| 1506 |
+
"tokens/train_per_sec_per_gpu": 11.23,
|
| 1507 |
+
"tokens/trainable": 214498
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.14347392892726668,
|
| 1511 |
+
"grad_norm": 0.2580498158931732,
|
| 1512 |
+
"learning_rate": 4.83504027183137e-05,
|
| 1513 |
+
"loss": 0.2704671323299408,
|
| 1514 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1515 |
+
"memory/max_active (GiB)": 75.56,
|
| 1516 |
+
"memory/max_allocated (GiB)": 75.56,
|
| 1517 |
+
"ppl": 1.31058,
|
| 1518 |
+
"step": 108,
|
| 1519 |
+
"tokens/total": 25867840,
|
| 1520 |
+
"tokens/train_per_sec_per_gpu": 11.28,
|
| 1521 |
+
"tokens/trainable": 216425
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.14480239123214877,
|
| 1525 |
+
"grad_norm": 0.29200857877731323,
|
| 1526 |
+
"learning_rate": 4.8311805735108894e-05,
|
| 1527 |
+
"loss": 0.33830636739730835,
|
| 1528 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1529 |
+
"memory/max_active (GiB)": 76.02,
|
| 1530 |
+
"memory/max_allocated (GiB)": 76.02,
|
| 1531 |
+
"ppl": 1.40257,
|
| 1532 |
+
"step": 109,
|
| 1533 |
+
"tokens/total": 26134080,
|
| 1534 |
+
"tokens/train_per_sec_per_gpu": 10.53,
|
| 1535 |
+
"tokens/trainable": 218302
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.1461308535370309,
|
| 1539 |
+
"grad_norm": 0.2611442804336548,
|
| 1540 |
+
"learning_rate": 4.827277818540251e-05,
|
| 1541 |
+
"loss": 0.2801956534385681,
|
| 1542 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1543 |
+
"memory/max_active (GiB)": 75.74,
|
| 1544 |
+
"memory/max_allocated (GiB)": 75.74,
|
| 1545 |
+
"ppl": 1.32339,
|
| 1546 |
+
"step": 110,
|
| 1547 |
+
"tokens/total": 26382912,
|
| 1548 |
+
"tokens/train_per_sec_per_gpu": 12.55,
|
| 1549 |
+
"tokens/trainable": 220213
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.147459315841913,
|
| 1553 |
+
"grad_norm": 0.31871896982192993,
|
| 1554 |
+
"learning_rate": 4.8233320790029106e-05,
|
| 1555 |
+
"loss": 0.295947402715683,
|
| 1556 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1557 |
+
"memory/max_active (GiB)": 73.9,
|
| 1558 |
+
"memory/max_allocated (GiB)": 73.9,
|
| 1559 |
+
"ppl": 1.3444,
|
| 1560 |
+
"step": 111,
|
| 1561 |
+
"tokens/total": 26601472,
|
| 1562 |
+
"tokens/train_per_sec_per_gpu": 12.96,
|
| 1563 |
+
"tokens/trainable": 222033
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.14878777814679509,
|
| 1567 |
+
"grad_norm": 0.2990409731864929,
|
| 1568 |
+
"learning_rate": 4.819343427776247e-05,
|
| 1569 |
+
"loss": 0.30516964197158813,
|
| 1570 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1571 |
+
"memory/max_active (GiB)": 73.84,
|
| 1572 |
+
"memory/max_allocated (GiB)": 73.84,
|
| 1573 |
+
"ppl": 1.35686,
|
| 1574 |
+
"step": 112,
|
| 1575 |
+
"tokens/total": 26833216,
|
| 1576 |
+
"tokens/train_per_sec_per_gpu": 12.57,
|
| 1577 |
+
"tokens/trainable": 224044
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.15011624045167718,
|
| 1581 |
+
"grad_norm": 0.33460795879364014,
|
| 1582 |
+
"learning_rate": 4.815311938530211e-05,
|
| 1583 |
+
"loss": 0.3858802914619446,
|
| 1584 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1585 |
+
"memory/max_active (GiB)": 76.4,
|
| 1586 |
+
"memory/max_allocated (GiB)": 76.4,
|
| 1587 |
+
"ppl": 1.47091,
|
| 1588 |
+
"step": 113,
|
| 1589 |
+
"tokens/total": 27077120,
|
| 1590 |
+
"tokens/train_per_sec_per_gpu": 12.46,
|
| 1591 |
+
"tokens/trainable": 226182
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.15144470275655927,
|
| 1595 |
+
"grad_norm": 0.2794348895549774,
|
| 1596 |
+
"learning_rate": 4.8112376857259696e-05,
|
| 1597 |
+
"loss": 0.2724725008010864,
|
| 1598 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1599 |
+
"memory/max_active (GiB)": 73.79,
|
| 1600 |
+
"memory/max_allocated (GiB)": 73.79,
|
| 1601 |
+
"ppl": 1.31321,
|
| 1602 |
+
"step": 114,
|
| 1603 |
+
"tokens/total": 27304640,
|
| 1604 |
+
"tokens/train_per_sec_per_gpu": 14.11,
|
| 1605 |
+
"tokens/trainable": 227959
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.15277316506144137,
|
| 1609 |
+
"grad_norm": 0.29718488454818726,
|
| 1610 |
+
"learning_rate": 4.8071207446145244e-05,
|
| 1611 |
+
"loss": 0.2983158528804779,
|
| 1612 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1613 |
+
"memory/max_active (GiB)": 73.05,
|
| 1614 |
+
"memory/max_allocated (GiB)": 73.05,
|
| 1615 |
+
"ppl": 1.34759,
|
| 1616 |
+
"step": 115,
|
| 1617 |
+
"tokens/total": 27544512,
|
| 1618 |
+
"tokens/train_per_sec_per_gpu": 12.63,
|
| 1619 |
+
"tokens/trainable": 230054
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.1541016273663235,
|
| 1623 |
+
"grad_norm": 0.25398096442222595,
|
| 1624 |
+
"learning_rate": 4.802961191235329e-05,
|
| 1625 |
+
"loss": 0.25493133068084717,
|
| 1626 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1627 |
+
"memory/max_active (GiB)": 74.14,
|
| 1628 |
+
"memory/max_allocated (GiB)": 74.14,
|
| 1629 |
+
"ppl": 1.29037,
|
| 1630 |
+
"step": 116,
|
| 1631 |
+
"tokens/total": 27787584,
|
| 1632 |
+
"tokens/train_per_sec_per_gpu": 11.73,
|
| 1633 |
+
"tokens/trainable": 231975
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.15543008967120558,
|
| 1637 |
+
"grad_norm": 0.5715415477752686,
|
| 1638 |
+
"learning_rate": 4.798759102414878e-05,
|
| 1639 |
+
"loss": 0.327411413192749,
|
| 1640 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1641 |
+
"memory/max_active (GiB)": 74.66,
|
| 1642 |
+
"memory/max_allocated (GiB)": 74.66,
|
| 1643 |
+
"ppl": 1.38737,
|
| 1644 |
+
"step": 117,
|
| 1645 |
+
"tokens/total": 28038336,
|
| 1646 |
+
"tokens/train_per_sec_per_gpu": 13.38,
|
| 1647 |
+
"tokens/trainable": 233982
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.15675855197608768,
|
| 1651 |
+
"grad_norm": 0.3132580518722534,
|
| 1652 |
+
"learning_rate": 4.794514555765293e-05,
|
| 1653 |
+
"loss": 0.334320604801178,
|
| 1654 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1655 |
+
"memory/max_active (GiB)": 71.85,
|
| 1656 |
+
"memory/max_allocated (GiB)": 71.85,
|
| 1657 |
+
"ppl": 1.39699,
|
| 1658 |
+
"step": 118,
|
| 1659 |
+
"tokens/total": 28277568,
|
| 1660 |
+
"tokens/train_per_sec_per_gpu": 13.47,
|
| 1661 |
+
"tokens/trainable": 235995
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.15808701428096977,
|
| 1665 |
+
"grad_norm": 0.3772391378879547,
|
| 1666 |
+
"learning_rate": 4.790227629682885e-05,
|
| 1667 |
+
"loss": 0.3174671530723572,
|
| 1668 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1669 |
+
"memory/max_active (GiB)": 75.45,
|
| 1670 |
+
"memory/max_allocated (GiB)": 75.45,
|
| 1671 |
+
"ppl": 1.37364,
|
| 1672 |
+
"step": 119,
|
| 1673 |
+
"tokens/total": 28516096,
|
| 1674 |
+
"tokens/train_per_sec_per_gpu": 13.49,
|
| 1675 |
+
"tokens/trainable": 238010
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.15941547658585187,
|
| 1679 |
+
"grad_norm": 0.39647936820983887,
|
| 1680 |
+
"learning_rate": 4.785898403346711e-05,
|
| 1681 |
+
"loss": 0.36018556356430054,
|
| 1682 |
+
"memory/device_reserved (GiB)": 77.53,
|
| 1683 |
+
"memory/max_active (GiB)": 72.56,
|
| 1684 |
+
"memory/max_allocated (GiB)": 72.56,
|
| 1685 |
+
"ppl": 1.4336,
|
| 1686 |
+
"step": 120,
|
| 1687 |
+
"tokens/total": 28744448,
|
| 1688 |
+
"tokens/train_per_sec_per_gpu": 14.59,
|
| 1689 |
+
"tokens/trainable": 240127
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.16074393889073396,
|
| 1693 |
+
"grad_norm": 0.24146312475204468,
|
| 1694 |
+
"learning_rate": 4.7815269567171084e-05,
|
| 1695 |
+
"loss": 0.24168699979782104,
|
| 1696 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1697 |
+
"memory/max_active (GiB)": 76.84,
|
| 1698 |
+
"memory/max_allocated (GiB)": 76.84,
|
| 1699 |
+
"ppl": 1.2734,
|
| 1700 |
+
"step": 121,
|
| 1701 |
+
"tokens/total": 29015616,
|
| 1702 |
+
"tokens/train_per_sec_per_gpu": 10.19,
|
| 1703 |
+
"tokens/trainable": 242077
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.16207240119561608,
|
| 1707 |
+
"grad_norm": 0.2935802936553955,
|
| 1708 |
+
"learning_rate": 4.777113370534216e-05,
|
| 1709 |
+
"loss": 0.3234092593193054,
|
| 1710 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1711 |
+
"memory/max_active (GiB)": 71.77,
|
| 1712 |
+
"memory/max_allocated (GiB)": 71.77,
|
| 1713 |
+
"ppl": 1.38183,
|
| 1714 |
+
"step": 122,
|
| 1715 |
+
"tokens/total": 29278976,
|
| 1716 |
+
"tokens/train_per_sec_per_gpu": 13.4,
|
| 1717 |
+
"tokens/trainable": 244412
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.16340086350049818,
|
| 1721 |
+
"grad_norm": 0.2624473571777344,
|
| 1722 |
+
"learning_rate": 4.7726577263164905e-05,
|
| 1723 |
+
"loss": 0.2843772768974304,
|
| 1724 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1725 |
+
"memory/max_active (GiB)": 71.2,
|
| 1726 |
+
"memory/max_allocated (GiB)": 71.2,
|
| 1727 |
+
"ppl": 1.32893,
|
| 1728 |
+
"step": 123,
|
| 1729 |
+
"tokens/total": 29514048,
|
| 1730 |
+
"tokens/train_per_sec_per_gpu": 11.68,
|
| 1731 |
+
"tokens/trainable": 246305
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.16472932580538027,
|
| 1735 |
+
"grad_norm": 0.2984730303287506,
|
| 1736 |
+
"learning_rate": 4.768160106359191e-05,
|
| 1737 |
+
"loss": 0.3387468457221985,
|
| 1738 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1739 |
+
"memory/max_active (GiB)": 74.79,
|
| 1740 |
+
"memory/max_allocated (GiB)": 74.79,
|
| 1741 |
+
"ppl": 1.40319,
|
| 1742 |
+
"step": 124,
|
| 1743 |
+
"tokens/total": 29758656,
|
| 1744 |
+
"tokens/train_per_sec_per_gpu": 14.46,
|
| 1745 |
+
"tokens/trainable": 248446
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.16605778811026237,
|
| 1749 |
+
"grad_norm": 0.27655282616615295,
|
| 1750 |
+
"learning_rate": 4.763620593732867e-05,
|
| 1751 |
+
"loss": 0.3312479853630066,
|
| 1752 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1753 |
+
"memory/max_active (GiB)": 76.29,
|
| 1754 |
+
"memory/max_allocated (GiB)": 76.29,
|
| 1755 |
+
"ppl": 1.39271,
|
| 1756 |
+
"step": 125,
|
| 1757 |
+
"tokens/total": 30048000,
|
| 1758 |
+
"tokens/train_per_sec_per_gpu": 10.94,
|
| 1759 |
+
"tokens/trainable": 250450
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.16738625041514446,
|
| 1763 |
+
"grad_norm": 0.3455372452735901,
|
| 1764 |
+
"learning_rate": 4.75903927228182e-05,
|
| 1765 |
+
"loss": 0.3507857918739319,
|
| 1766 |
+
"memory/device_reserved (GiB)": 77.73,
|
| 1767 |
+
"memory/max_active (GiB)": 72.86,
|
| 1768 |
+
"memory/max_allocated (GiB)": 72.86,
|
| 1769 |
+
"ppl": 1.42018,
|
| 1770 |
+
"step": 126,
|
| 1771 |
+
"tokens/total": 30278784,
|
| 1772 |
+
"tokens/train_per_sec_per_gpu": 9.1,
|
| 1773 |
+
"tokens/trainable": 252279
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.16871471272002658,
|
| 1777 |
+
"grad_norm": 0.2768387496471405,
|
| 1778 |
+
"learning_rate": 4.7544162266225546e-05,
|
| 1779 |
+
"loss": 0.3144207000732422,
|
| 1780 |
+
"memory/device_reserved (GiB)": 75.11,
|
| 1781 |
+
"memory/max_active (GiB)": 74.47,
|
| 1782 |
+
"memory/max_allocated (GiB)": 74.47,
|
| 1783 |
+
"ppl": 1.36947,
|
| 1784 |
+
"step": 127,
|
| 1785 |
+
"tokens/total": 30532800,
|
| 1786 |
+
"tokens/train_per_sec_per_gpu": 13.95,
|
| 1787 |
+
"tokens/trainable": 254541
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.17004317502490868,
|
| 1791 |
+
"grad_norm": 0.2863609790802002,
|
| 1792 |
+
"learning_rate": 4.7497515421422207e-05,
|
| 1793 |
+
"loss": 0.3234996199607849,
|
| 1794 |
+
"memory/device_reserved (GiB)": 77.65,
|
| 1795 |
+
"memory/max_active (GiB)": 76.75,
|
| 1796 |
+
"memory/max_allocated (GiB)": 76.75,
|
| 1797 |
+
"ppl": 1.38196,
|
| 1798 |
+
"step": 128,
|
| 1799 |
+
"tokens/total": 30798336,
|
| 1800 |
+
"tokens/train_per_sec_per_gpu": 12.97,
|
| 1801 |
+
"tokens/trainable": 256801
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.17137163732979077,
|
| 1805 |
+
"grad_norm": 0.32456719875335693,
|
| 1806 |
+
"learning_rate": 4.745045304997029e-05,
|
| 1807 |
+
"loss": 0.3574495315551758,
|
| 1808 |
+
"memory/device_reserved (GiB)": 77.65,
|
| 1809 |
+
"memory/max_active (GiB)": 74.69,
|
| 1810 |
+
"memory/max_allocated (GiB)": 74.69,
|
| 1811 |
+
"ppl": 1.42968,
|
| 1812 |
+
"step": 129,
|
| 1813 |
+
"tokens/total": 31032448,
|
| 1814 |
+
"tokens/train_per_sec_per_gpu": 11.0,
|
| 1815 |
+
"tokens/trainable": 258757
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.17270009963467287,
|
| 1819 |
+
"grad_norm": 0.31199607253074646,
|
| 1820 |
+
"learning_rate": 4.7402976021106646e-05,
|
| 1821 |
+
"loss": 0.4081224501132965,
|
| 1822 |
+
"memory/device_reserved (GiB)": 77.65,
|
| 1823 |
+
"memory/max_active (GiB)": 74.93,
|
| 1824 |
+
"memory/max_allocated (GiB)": 74.93,
|
| 1825 |
+
"ppl": 1.50399,
|
| 1826 |
+
"step": 130,
|
| 1827 |
+
"tokens/total": 31266880,
|
| 1828 |
+
"tokens/train_per_sec_per_gpu": 13.45,
|
| 1829 |
+
"tokens/trainable": 260916
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.17402856193955496,
|
| 1833 |
+
"grad_norm": 0.281172513961792,
|
| 1834 |
+
"learning_rate": 4.735508521172681e-05,
|
| 1835 |
+
"loss": 0.288382887840271,
|
| 1836 |
+
"memory/device_reserved (GiB)": 77.65,
|
| 1837 |
+
"memory/max_active (GiB)": 74.41,
|
| 1838 |
+
"memory/max_allocated (GiB)": 74.41,
|
| 1839 |
+
"ppl": 1.33427,
|
| 1840 |
+
"step": 131,
|
| 1841 |
+
"tokens/total": 31526016,
|
| 1842 |
+
"tokens/train_per_sec_per_gpu": 12.06,
|
| 1843 |
+
"tokens/trainable": 263039
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.17535702424443705,
|
| 1847 |
+
"grad_norm": 0.32762470841407776,
|
| 1848 |
+
"learning_rate": 4.730678150636879e-05,
|
| 1849 |
+
"loss": 0.334347128868103,
|
| 1850 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1851 |
+
"memory/max_active (GiB)": 76.81,
|
| 1852 |
+
"memory/max_allocated (GiB)": 76.81,
|
| 1853 |
+
"ppl": 1.39703,
|
| 1854 |
+
"step": 132,
|
| 1855 |
+
"tokens/total": 31792512,
|
| 1856 |
+
"tokens/train_per_sec_per_gpu": 11.84,
|
| 1857 |
+
"tokens/trainable": 265042
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.17668548654931918,
|
| 1861 |
+
"grad_norm": 0.2732056975364685,
|
| 1862 |
+
"learning_rate": 4.7258065797196746e-05,
|
| 1863 |
+
"loss": 0.2946048378944397,
|
| 1864 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1865 |
+
"memory/max_active (GiB)": 75.15,
|
| 1866 |
+
"memory/max_allocated (GiB)": 75.15,
|
| 1867 |
+
"ppl": 1.3426,
|
| 1868 |
+
"step": 133,
|
| 1869 |
+
"tokens/total": 32026880,
|
| 1870 |
+
"tokens/train_per_sec_per_gpu": 13.67,
|
| 1871 |
+
"tokens/trainable": 267150
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.17801394885420127,
|
| 1875 |
+
"grad_norm": 0.25782883167266846,
|
| 1876 |
+
"learning_rate": 4.720893898398448e-05,
|
| 1877 |
+
"loss": 0.2663328945636749,
|
| 1878 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1879 |
+
"memory/max_active (GiB)": 75.94,
|
| 1880 |
+
"memory/max_allocated (GiB)": 75.94,
|
| 1881 |
+
"ppl": 1.30517,
|
| 1882 |
+
"step": 134,
|
| 1883 |
+
"tokens/total": 32308160,
|
| 1884 |
+
"tokens/train_per_sec_per_gpu": 11.08,
|
| 1885 |
+
"tokens/trainable": 269279
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.17934241115908336,
|
| 1889 |
+
"grad_norm": 0.342846542596817,
|
| 1890 |
+
"learning_rate": 4.715940197409887e-05,
|
| 1891 |
+
"loss": 0.2981494069099426,
|
| 1892 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1893 |
+
"memory/max_active (GiB)": 70.77,
|
| 1894 |
+
"memory/max_allocated (GiB)": 70.77,
|
| 1895 |
+
"ppl": 1.34736,
|
| 1896 |
+
"step": 135,
|
| 1897 |
+
"tokens/total": 32536512,
|
| 1898 |
+
"tokens/train_per_sec_per_gpu": 10.64,
|
| 1899 |
+
"tokens/trainable": 271049
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.18067087346396546,
|
| 1903 |
+
"grad_norm": 0.3146810829639435,
|
| 1904 |
+
"learning_rate": 4.710945568248306e-05,
|
| 1905 |
+
"loss": 0.3311242461204529,
|
| 1906 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1907 |
+
"memory/max_active (GiB)": 76.37,
|
| 1908 |
+
"memory/max_allocated (GiB)": 76.37,
|
| 1909 |
+
"ppl": 1.39253,
|
| 1910 |
+
"step": 136,
|
| 1911 |
+
"tokens/total": 32801920,
|
| 1912 |
+
"tokens/train_per_sec_per_gpu": 11.55,
|
| 1913 |
+
"tokens/trainable": 272970
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.18199933576884755,
|
| 1917 |
+
"grad_norm": 0.358317106962204,
|
| 1918 |
+
"learning_rate": 4.705910103163961e-05,
|
| 1919 |
+
"loss": 0.28497856855392456,
|
| 1920 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1921 |
+
"memory/max_active (GiB)": 73.76,
|
| 1922 |
+
"memory/max_allocated (GiB)": 73.76,
|
| 1923 |
+
"ppl": 1.32973,
|
| 1924 |
+
"step": 137,
|
| 1925 |
+
"tokens/total": 33045184,
|
| 1926 |
+
"tokens/train_per_sec_per_gpu": 13.69,
|
| 1927 |
+
"tokens/trainable": 275029
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.18332779807372965,
|
| 1931 |
+
"grad_norm": 0.27245497703552246,
|
| 1932 |
+
"learning_rate": 4.7008338951613394e-05,
|
| 1933 |
+
"loss": 0.28498828411102295,
|
| 1934 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1935 |
+
"memory/max_active (GiB)": 75.31,
|
| 1936 |
+
"memory/max_allocated (GiB)": 75.31,
|
| 1937 |
+
"ppl": 1.32975,
|
| 1938 |
+
"step": 138,
|
| 1939 |
+
"tokens/total": 33309824,
|
| 1940 |
+
"tokens/train_per_sec_per_gpu": 13.06,
|
| 1941 |
+
"tokens/trainable": 277152
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.18465626037861177,
|
| 1945 |
+
"grad_norm": 0.29116177558898926,
|
| 1946 |
+
"learning_rate": 4.695717037997448e-05,
|
| 1947 |
+
"loss": 0.275570273399353,
|
| 1948 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1949 |
+
"memory/max_active (GiB)": 74.06,
|
| 1950 |
+
"memory/max_allocated (GiB)": 74.06,
|
| 1951 |
+
"ppl": 1.31728,
|
| 1952 |
+
"step": 139,
|
| 1953 |
+
"tokens/total": 33522560,
|
| 1954 |
+
"tokens/train_per_sec_per_gpu": 13.63,
|
| 1955 |
+
"tokens/trainable": 278902
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.18598472268349386,
|
| 1959 |
+
"grad_norm": 0.2572404742240906,
|
| 1960 |
+
"learning_rate": 4.69055962618008e-05,
|
| 1961 |
+
"loss": 0.2584611177444458,
|
| 1962 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1963 |
+
"memory/max_active (GiB)": 75.82,
|
| 1964 |
+
"memory/max_allocated (GiB)": 75.82,
|
| 1965 |
+
"ppl": 1.29494,
|
| 1966 |
+
"step": 140,
|
| 1967 |
+
"tokens/total": 33747968,
|
| 1968 |
+
"tokens/train_per_sec_per_gpu": 14.16,
|
| 1969 |
+
"tokens/trainable": 280769
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.18731318498837596,
|
| 1973 |
+
"grad_norm": 0.2903401255607605,
|
| 1974 |
+
"learning_rate": 4.685361754966065e-05,
|
| 1975 |
+
"loss": 0.3318089246749878,
|
| 1976 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1977 |
+
"memory/max_active (GiB)": 75.07,
|
| 1978 |
+
"memory/max_allocated (GiB)": 75.07,
|
| 1979 |
+
"ppl": 1.39349,
|
| 1980 |
+
"step": 141,
|
| 1981 |
+
"tokens/total": 33982976,
|
| 1982 |
+
"tokens/train_per_sec_per_gpu": 16.53,
|
| 1983 |
+
"tokens/trainable": 282960
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.18864164729325805,
|
| 1987 |
+
"grad_norm": 0.28281405568122864,
|
| 1988 |
+
"learning_rate": 4.6801235203595195e-05,
|
| 1989 |
+
"loss": 0.26338034868240356,
|
| 1990 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 1991 |
+
"memory/max_active (GiB)": 74.9,
|
| 1992 |
+
"memory/max_allocated (GiB)": 74.9,
|
| 1993 |
+
"ppl": 1.30132,
|
| 1994 |
+
"step": 142,
|
| 1995 |
+
"tokens/total": 34235328,
|
| 1996 |
+
"tokens/train_per_sec_per_gpu": 12.81,
|
| 1997 |
+
"tokens/trainable": 285014
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.18997010959814015,
|
| 2001 |
+
"grad_norm": 0.35500800609588623,
|
| 2002 |
+
"learning_rate": 4.67484501911006e-05,
|
| 2003 |
+
"loss": 0.40566298365592957,
|
| 2004 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2005 |
+
"memory/max_active (GiB)": 74.06,
|
| 2006 |
+
"memory/max_allocated (GiB)": 74.06,
|
| 2007 |
+
"ppl": 1.5003,
|
| 2008 |
+
"step": 143,
|
| 2009 |
+
"tokens/total": 34485952,
|
| 2010 |
+
"tokens/train_per_sec_per_gpu": 13.65,
|
| 2011 |
+
"tokens/trainable": 287043
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.19129857190302224,
|
| 2015 |
+
"grad_norm": 0.2855430245399475,
|
| 2016 |
+
"learning_rate": 4.669526348711028e-05,
|
| 2017 |
+
"loss": 0.2761286497116089,
|
| 2018 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2019 |
+
"memory/max_active (GiB)": 73.19,
|
| 2020 |
+
"memory/max_allocated (GiB)": 73.19,
|
| 2021 |
+
"ppl": 1.31802,
|
| 2022 |
+
"step": 144,
|
| 2023 |
+
"tokens/total": 34707328,
|
| 2024 |
+
"tokens/train_per_sec_per_gpu": 16.14,
|
| 2025 |
+
"tokens/trainable": 289319
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.19262703420790436,
|
| 2029 |
+
"grad_norm": 0.26699694991111755,
|
| 2030 |
+
"learning_rate": 4.664167607397686e-05,
|
| 2031 |
+
"loss": 0.23146027326583862,
|
| 2032 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2033 |
+
"memory/max_active (GiB)": 74.28,
|
| 2034 |
+
"memory/max_allocated (GiB)": 74.28,
|
| 2035 |
+
"ppl": 1.26044,
|
| 2036 |
+
"step": 145,
|
| 2037 |
+
"tokens/total": 34971200,
|
| 2038 |
+
"tokens/train_per_sec_per_gpu": 10.74,
|
| 2039 |
+
"tokens/trainable": 291326
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.19395549651278646,
|
| 2043 |
+
"grad_norm": 0.32218310236930847,
|
| 2044 |
+
"learning_rate": 4.658768894145397e-05,
|
| 2045 |
+
"loss": 0.32315659523010254,
|
| 2046 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2047 |
+
"memory/max_active (GiB)": 76.67,
|
| 2048 |
+
"memory/max_allocated (GiB)": 76.67,
|
| 2049 |
+
"ppl": 1.38148,
|
| 2050 |
+
"step": 146,
|
| 2051 |
+
"tokens/total": 35194240,
|
| 2052 |
+
"tokens/train_per_sec_per_gpu": 15.46,
|
| 2053 |
+
"tokens/trainable": 293573
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.19528395881766855,
|
| 2057 |
+
"grad_norm": 0.3315838575363159,
|
| 2058 |
+
"learning_rate": 4.653330308667805e-05,
|
| 2059 |
+
"loss": 0.2897493243217468,
|
| 2060 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2061 |
+
"memory/max_active (GiB)": 73.57,
|
| 2062 |
+
"memory/max_allocated (GiB)": 73.57,
|
| 2063 |
+
"ppl": 1.33609,
|
| 2064 |
+
"step": 147,
|
| 2065 |
+
"tokens/total": 35438912,
|
| 2066 |
+
"tokens/train_per_sec_per_gpu": 13.66,
|
| 2067 |
+
"tokens/trainable": 295640
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.19661242112255065,
|
| 2071 |
+
"grad_norm": 0.28656816482543945,
|
| 2072 |
+
"learning_rate": 4.64785195141499e-05,
|
| 2073 |
+
"loss": 0.3137304186820984,
|
| 2074 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2075 |
+
"memory/max_active (GiB)": 73.57,
|
| 2076 |
+
"memory/max_allocated (GiB)": 73.57,
|
| 2077 |
+
"ppl": 1.36852,
|
| 2078 |
+
"step": 148,
|
| 2079 |
+
"tokens/total": 35660288,
|
| 2080 |
+
"tokens/train_per_sec_per_gpu": 14.06,
|
| 2081 |
+
"tokens/trainable": 297574
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.19794088342743274,
|
| 2085 |
+
"grad_norm": 0.29270216822624207,
|
| 2086 |
+
"learning_rate": 4.64233392357161e-05,
|
| 2087 |
+
"loss": 0.2963527739048004,
|
| 2088 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2089 |
+
"memory/max_active (GiB)": 75.34,
|
| 2090 |
+
"memory/max_allocated (GiB)": 75.34,
|
| 2091 |
+
"ppl": 1.34494,
|
| 2092 |
+
"step": 149,
|
| 2093 |
+
"tokens/total": 35924928,
|
| 2094 |
+
"tokens/train_per_sec_per_gpu": 11.77,
|
| 2095 |
+
"tokens/trainable": 299725
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.19926934573231483,
|
| 2099 |
+
"grad_norm": 0.3433849811553955,
|
| 2100 |
+
"learning_rate": 4.636776327055035e-05,
|
| 2101 |
+
"loss": 0.317320317029953,
|
| 2102 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2103 |
+
"memory/max_active (GiB)": 75.83,
|
| 2104 |
+
"memory/max_allocated (GiB)": 75.83,
|
| 2105 |
+
"ppl": 1.37344,
|
| 2106 |
+
"step": 150,
|
| 2107 |
+
"tokens/total": 36142656,
|
| 2108 |
+
"tokens/train_per_sec_per_gpu": 12.91,
|
| 2109 |
+
"tokens/trainable": 301695
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.20059780803719696,
|
| 2113 |
+
"grad_norm": 0.31213390827178955,
|
| 2114 |
+
"learning_rate": 4.6311792645134666e-05,
|
| 2115 |
+
"loss": 0.32360923290252686,
|
| 2116 |
+
"memory/device_reserved (GiB)": 77.77,
|
| 2117 |
+
"memory/max_active (GiB)": 76.75,
|
| 2118 |
+
"memory/max_allocated (GiB)": 76.75,
|
| 2119 |
+
"ppl": 1.38211,
|
| 2120 |
+
"step": 151,
|
| 2121 |
+
"tokens/total": 36373632,
|
| 2122 |
+
"tokens/train_per_sec_per_gpu": 9.87,
|
| 2123 |
+
"tokens/trainable": 303672
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.20192627034207905,
|
| 2127 |
+
"grad_norm": 0.3097369074821472,
|
| 2128 |
+
"learning_rate": 4.625542839324036e-05,
|
| 2129 |
+
"loss": 0.3049156963825226,
|
| 2130 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2131 |
+
"memory/max_active (GiB)": 74.93,
|
| 2132 |
+
"memory/max_allocated (GiB)": 74.93,
|
| 2133 |
+
"ppl": 1.35651,
|
| 2134 |
+
"step": 152,
|
| 2135 |
+
"tokens/total": 36624000,
|
| 2136 |
+
"tokens/train_per_sec_per_gpu": 13.25,
|
| 2137 |
+
"tokens/trainable": 305735
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.20325473264696114,
|
| 2141 |
+
"grad_norm": 0.30076032876968384,
|
| 2142 |
+
"learning_rate": 4.6198671555909e-05,
|
| 2143 |
+
"loss": 0.32292836904525757,
|
| 2144 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2145 |
+
"memory/max_active (GiB)": 74.6,
|
| 2146 |
+
"memory/max_allocated (GiB)": 74.6,
|
| 2147 |
+
"ppl": 1.38117,
|
| 2148 |
+
"step": 153,
|
| 2149 |
+
"tokens/total": 36853696,
|
| 2150 |
+
"tokens/train_per_sec_per_gpu": 13.61,
|
| 2151 |
+
"tokens/trainable": 307623
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.20458319495184324,
|
| 2155 |
+
"grad_norm": 0.3042234182357788,
|
| 2156 |
+
"learning_rate": 4.6141523181433165e-05,
|
| 2157 |
+
"loss": 0.34925395250320435,
|
| 2158 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2159 |
+
"memory/max_active (GiB)": 75.88,
|
| 2160 |
+
"memory/max_allocated (GiB)": 75.88,
|
| 2161 |
+
"ppl": 1.41801,
|
| 2162 |
+
"step": 154,
|
| 2163 |
+
"tokens/total": 37093312,
|
| 2164 |
+
"tokens/train_per_sec_per_gpu": 14.79,
|
| 2165 |
+
"tokens/trainable": 309875
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.20591165725672533,
|
| 2169 |
+
"grad_norm": 0.25751274824142456,
|
| 2170 |
+
"learning_rate": 4.608398432533708e-05,
|
| 2171 |
+
"loss": 0.22095003724098206,
|
| 2172 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2173 |
+
"memory/max_active (GiB)": 75.39,
|
| 2174 |
+
"memory/max_allocated (GiB)": 75.39,
|
| 2175 |
+
"ppl": 1.24726,
|
| 2176 |
+
"step": 155,
|
| 2177 |
+
"tokens/total": 37312896,
|
| 2178 |
+
"tokens/train_per_sec_per_gpu": 11.13,
|
| 2179 |
+
"tokens/trainable": 311609
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.20724011956160743,
|
| 2183 |
+
"grad_norm": 0.2809881567955017,
|
| 2184 |
+
"learning_rate": 4.602605605035713e-05,
|
| 2185 |
+
"loss": 0.3040582835674286,
|
| 2186 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2187 |
+
"memory/max_active (GiB)": 73.62,
|
| 2188 |
+
"memory/max_allocated (GiB)": 73.62,
|
| 2189 |
+
"ppl": 1.35535,
|
| 2190 |
+
"step": 156,
|
| 2191 |
+
"tokens/total": 37581184,
|
| 2192 |
+
"tokens/train_per_sec_per_gpu": 11.32,
|
| 2193 |
+
"tokens/trainable": 313404
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.20856858186648955,
|
| 2197 |
+
"grad_norm": 0.2722752094268799,
|
| 2198 |
+
"learning_rate": 4.5967739426422193e-05,
|
| 2199 |
+
"loss": 0.24645733833312988,
|
| 2200 |
+
"memory/device_reserved (GiB)": 77.43,
|
| 2201 |
+
"memory/max_active (GiB)": 74.28,
|
| 2202 |
+
"memory/max_allocated (GiB)": 74.28,
|
| 2203 |
+
"ppl": 1.27948,
|
| 2204 |
+
"step": 157,
|
| 2205 |
+
"tokens/total": 37820352,
|
| 2206 |
+
"tokens/train_per_sec_per_gpu": 11.93,
|
| 2207 |
+
"tokens/trainable": 315316
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.20989704417137164,
|
| 2211 |
+
"grad_norm": 0.2881944477558136,
|
| 2212 |
+
"learning_rate": 4.590903553063396e-05,
|
| 2213 |
+
"loss": 0.29333794116973877,
|
| 2214 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2215 |
+
"memory/max_active (GiB)": 76.67,
|
| 2216 |
+
"memory/max_allocated (GiB)": 76.67,
|
| 2217 |
+
"ppl": 1.3409,
|
| 2218 |
+
"step": 158,
|
| 2219 |
+
"tokens/total": 38062400,
|
| 2220 |
+
"tokens/train_per_sec_per_gpu": 13.87,
|
| 2221 |
+
"tokens/trainable": 317431
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.21122550647625374,
|
| 2225 |
+
"grad_norm": 0.3498353064060211,
|
| 2226 |
+
"learning_rate": 4.584994544724695e-05,
|
| 2227 |
+
"loss": 0.2311086356639862,
|
| 2228 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2229 |
+
"memory/max_active (GiB)": 72.75,
|
| 2230 |
+
"memory/max_allocated (GiB)": 72.75,
|
| 2231 |
+
"ppl": 1.26,
|
| 2232 |
+
"step": 159,
|
| 2233 |
+
"tokens/total": 38270528,
|
| 2234 |
+
"tokens/train_per_sec_per_gpu": 14.7,
|
| 2235 |
+
"tokens/trainable": 319251
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.21255396878113583,
|
| 2239 |
+
"grad_norm": 0.3815538287162781,
|
| 2240 |
+
"learning_rate": 4.5790470267648546e-05,
|
| 2241 |
+
"loss": 0.3265290856361389,
|
| 2242 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2243 |
+
"memory/max_active (GiB)": 74.9,
|
| 2244 |
+
"memory/max_allocated (GiB)": 74.9,
|
| 2245 |
+
"ppl": 1.38615,
|
| 2246 |
+
"step": 160,
|
| 2247 |
+
"tokens/total": 38513600,
|
| 2248 |
+
"tokens/train_per_sec_per_gpu": 13.21,
|
| 2249 |
+
"tokens/trainable": 321202
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.21388243108601793,
|
| 2253 |
+
"grad_norm": 0.2811428904533386,
|
| 2254 |
+
"learning_rate": 4.573061109033879e-05,
|
| 2255 |
+
"loss": 0.28701144456863403,
|
| 2256 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2257 |
+
"memory/max_active (GiB)": 75.56,
|
| 2258 |
+
"memory/max_allocated (GiB)": 75.56,
|
| 2259 |
+
"ppl": 1.33244,
|
| 2260 |
+
"step": 161,
|
| 2261 |
+
"tokens/total": 38774720,
|
| 2262 |
+
"tokens/train_per_sec_per_gpu": 12.19,
|
| 2263 |
+
"tokens/trainable": 323209
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.21521089339090002,
|
| 2267 |
+
"grad_norm": 0.28593000769615173,
|
| 2268 |
+
"learning_rate": 4.5670369020910165e-05,
|
| 2269 |
+
"loss": 0.25745725631713867,
|
| 2270 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2271 |
+
"memory/max_active (GiB)": 76.24,
|
| 2272 |
+
"memory/max_allocated (GiB)": 76.24,
|
| 2273 |
+
"ppl": 1.29364,
|
| 2274 |
+
"step": 162,
|
| 2275 |
+
"tokens/total": 39011968,
|
| 2276 |
+
"tokens/train_per_sec_per_gpu": 13.34,
|
| 2277 |
+
"tokens/trainable": 325118
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.21653935569578214,
|
| 2281 |
+
"grad_norm": 0.2986576557159424,
|
| 2282 |
+
"learning_rate": 4.560974517202709e-05,
|
| 2283 |
+
"loss": 0.2792147994041443,
|
| 2284 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2285 |
+
"memory/max_active (GiB)": 76.26,
|
| 2286 |
+
"memory/max_allocated (GiB)": 76.26,
|
| 2287 |
+
"ppl": 1.32209,
|
| 2288 |
+
"step": 163,
|
| 2289 |
+
"tokens/total": 39264832,
|
| 2290 |
+
"tokens/train_per_sec_per_gpu": 10.73,
|
| 2291 |
+
"tokens/trainable": 326998
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.21786781800066424,
|
| 2295 |
+
"grad_norm": 0.2871696352958679,
|
| 2296 |
+
"learning_rate": 4.554874066340541e-05,
|
| 2297 |
+
"loss": 0.288088321685791,
|
| 2298 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2299 |
+
"memory/max_active (GiB)": 75.42,
|
| 2300 |
+
"memory/max_allocated (GiB)": 75.42,
|
| 2301 |
+
"ppl": 1.33388,
|
| 2302 |
+
"step": 164,
|
| 2303 |
+
"tokens/total": 39501440,
|
| 2304 |
+
"tokens/train_per_sec_per_gpu": 13.11,
|
| 2305 |
+
"tokens/trainable": 329049
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.21919628030554633,
|
| 2309 |
+
"grad_norm": 0.37666866183280945,
|
| 2310 |
+
"learning_rate": 4.548735662179174e-05,
|
| 2311 |
+
"loss": 0.34987354278564453,
|
| 2312 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2313 |
+
"memory/max_active (GiB)": 76.29,
|
| 2314 |
+
"memory/max_allocated (GiB)": 76.29,
|
| 2315 |
+
"ppl": 1.41889,
|
| 2316 |
+
"step": 165,
|
| 2317 |
+
"tokens/total": 39751360,
|
| 2318 |
+
"tokens/train_per_sec_per_gpu": 12.12,
|
| 2319 |
+
"tokens/trainable": 331007
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.22052474261042843,
|
| 2323 |
+
"grad_norm": 0.30536556243896484,
|
| 2324 |
+
"learning_rate": 4.542559418094261e-05,
|
| 2325 |
+
"loss": 0.27000105381011963,
|
| 2326 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2327 |
+
"memory/max_active (GiB)": 71.86,
|
| 2328 |
+
"memory/max_allocated (GiB)": 71.86,
|
| 2329 |
+
"ppl": 1.30997,
|
| 2330 |
+
"step": 166,
|
| 2331 |
+
"tokens/total": 39971904,
|
| 2332 |
+
"tokens/train_per_sec_per_gpu": 14.19,
|
| 2333 |
+
"tokens/trainable": 333019
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.22185320491531052,
|
| 2337 |
+
"grad_norm": 0.28819429874420166,
|
| 2338 |
+
"learning_rate": 4.536345448160355e-05,
|
| 2339 |
+
"loss": 0.26946187019348145,
|
| 2340 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2341 |
+
"memory/max_active (GiB)": 75.37,
|
| 2342 |
+
"memory/max_allocated (GiB)": 75.37,
|
| 2343 |
+
"ppl": 1.30926,
|
| 2344 |
+
"step": 167,
|
| 2345 |
+
"tokens/total": 40189632,
|
| 2346 |
+
"tokens/train_per_sec_per_gpu": 13.38,
|
| 2347 |
+
"tokens/trainable": 334772
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.22318166722019261,
|
| 2351 |
+
"grad_norm": 0.2997879981994629,
|
| 2352 |
+
"learning_rate": 4.5300938671487994e-05,
|
| 2353 |
+
"loss": 0.3346363306045532,
|
| 2354 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2355 |
+
"memory/max_active (GiB)": 74.55,
|
| 2356 |
+
"memory/max_allocated (GiB)": 74.55,
|
| 2357 |
+
"ppl": 1.39743,
|
| 2358 |
+
"step": 168,
|
| 2359 |
+
"tokens/total": 40410560,
|
| 2360 |
+
"tokens/train_per_sec_per_gpu": 14.8,
|
| 2361 |
+
"tokens/trainable": 336836
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.22451012952507474,
|
| 2365 |
+
"grad_norm": 0.347880482673645,
|
| 2366 |
+
"learning_rate": 4.5238047905256133e-05,
|
| 2367 |
+
"loss": 0.29474785923957825,
|
| 2368 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2369 |
+
"memory/max_active (GiB)": 74.17,
|
| 2370 |
+
"memory/max_allocated (GiB)": 74.17,
|
| 2371 |
+
"ppl": 1.34279,
|
| 2372 |
+
"step": 169,
|
| 2373 |
+
"tokens/total": 40637568,
|
| 2374 |
+
"tokens/train_per_sec_per_gpu": 12.59,
|
| 2375 |
+
"tokens/trainable": 338683
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.22583859182995683,
|
| 2379 |
+
"grad_norm": 0.3111642003059387,
|
| 2380 |
+
"learning_rate": 4.517478334449353e-05,
|
| 2381 |
+
"loss": 0.29106810688972473,
|
| 2382 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2383 |
+
"memory/max_active (GiB)": 74.85,
|
| 2384 |
+
"memory/max_allocated (GiB)": 74.85,
|
| 2385 |
+
"ppl": 1.33786,
|
| 2386 |
+
"step": 170,
|
| 2387 |
+
"tokens/total": 40868480,
|
| 2388 |
+
"tokens/train_per_sec_per_gpu": 14.64,
|
| 2389 |
+
"tokens/trainable": 340786
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.22716705413483892,
|
| 2393 |
+
"grad_norm": 0.31952518224716187,
|
| 2394 |
+
"learning_rate": 4.51111461576897e-05,
|
| 2395 |
+
"loss": 0.27892524003982544,
|
| 2396 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2397 |
+
"memory/max_active (GiB)": 73.84,
|
| 2398 |
+
"memory/max_allocated (GiB)": 73.84,
|
| 2399 |
+
"ppl": 1.32171,
|
| 2400 |
+
"step": 171,
|
| 2401 |
+
"tokens/total": 41112064,
|
| 2402 |
+
"tokens/train_per_sec_per_gpu": 13.66,
|
| 2403 |
+
"tokens/trainable": 342770
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.22849551643972102,
|
| 2407 |
+
"grad_norm": 0.3235322833061218,
|
| 2408 |
+
"learning_rate": 4.5047137520216525e-05,
|
| 2409 |
+
"loss": 0.2796081304550171,
|
| 2410 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2411 |
+
"memory/max_active (GiB)": 75.99,
|
| 2412 |
+
"memory/max_allocated (GiB)": 75.99,
|
| 2413 |
+
"ppl": 1.32261,
|
| 2414 |
+
"step": 172,
|
| 2415 |
+
"tokens/total": 41374592,
|
| 2416 |
+
"tokens/train_per_sec_per_gpu": 11.71,
|
| 2417 |
+
"tokens/trainable": 344707
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.2298239787446031,
|
| 2421 |
+
"grad_norm": 0.2968906760215759,
|
| 2422 |
+
"learning_rate": 4.4982758614306536e-05,
|
| 2423 |
+
"loss": 0.298882395029068,
|
| 2424 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2425 |
+
"memory/max_active (GiB)": 75.04,
|
| 2426 |
+
"memory/max_allocated (GiB)": 75.04,
|
| 2427 |
+
"ppl": 1.34835,
|
| 2428 |
+
"step": 173,
|
| 2429 |
+
"tokens/total": 41624000,
|
| 2430 |
+
"tokens/train_per_sec_per_gpu": 14.44,
|
| 2431 |
+
"tokens/trainable": 346771
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.2311524410494852,
|
| 2435 |
+
"grad_norm": 0.3555050492286682,
|
| 2436 |
+
"learning_rate": 4.491801062903108e-05,
|
| 2437 |
+
"loss": 0.3100724220275879,
|
| 2438 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2439 |
+
"memory/max_active (GiB)": 74.28,
|
| 2440 |
+
"memory/max_allocated (GiB)": 74.28,
|
| 2441 |
+
"ppl": 1.36352,
|
| 2442 |
+
"step": 174,
|
| 2443 |
+
"tokens/total": 41874048,
|
| 2444 |
+
"tokens/train_per_sec_per_gpu": 13.33,
|
| 2445 |
+
"tokens/trainable": 348815
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.23248090335436733,
|
| 2449 |
+
"grad_norm": 0.30196645855903625,
|
| 2450 |
+
"learning_rate": 4.4852894760278365e-05,
|
| 2451 |
+
"loss": 0.29532450437545776,
|
| 2452 |
+
"memory/device_reserved (GiB)": 77.51,
|
| 2453 |
+
"memory/max_active (GiB)": 71.06,
|
| 2454 |
+
"memory/max_allocated (GiB)": 71.06,
|
| 2455 |
+
"ppl": 1.34356,
|
| 2456 |
+
"step": 175,
|
| 2457 |
+
"tokens/total": 42112192,
|
| 2458 |
+
"tokens/train_per_sec_per_gpu": 11.1,
|
| 2459 |
+
"tokens/trainable": 350521
|
| 2460 |
+
}
|
| 2461 |
+
],
|
| 2462 |
+
"logging_steps": 1,
|
| 2463 |
+
"max_steps": 753,
|
| 2464 |
+
"num_input_tokens_seen": 0,
|
| 2465 |
+
"num_train_epochs": 1,
|
| 2466 |
+
"save_steps": 25,
|
| 2467 |
+
"stateful_callbacks": {
|
| 2468 |
+
"TrainerControl": {
|
| 2469 |
+
"args": {
|
| 2470 |
+
"should_epoch_stop": false,
|
| 2471 |
+
"should_evaluate": false,
|
| 2472 |
+
"should_log": false,
|
| 2473 |
+
"should_save": true,
|
| 2474 |
+
"should_training_stop": false
|
| 2475 |
+
},
|
| 2476 |
+
"attributes": {}
|
| 2477 |
+
}
|
| 2478 |
+
},
|
| 2479 |
+
"total_flos": 7.798370933936226e+18,
|
| 2480 |
+
"train_batch_size": 1,
|
| 2481 |
+
"trial_name": null,
|
| 2482 |
+
"trial_params": null
|
| 2483 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b976f66a095fd068f5cd7e9d460b7377c2040739c645602ed7ebc94358be660
|
| 3 |
+
size 6673
|