xyzzy
commited on
Commit
•
57c9597
1
Parent(s):
98bf8e1
a metanarrative begins to play out across the individual narratives of per-repository files. lonely, isolated commits are sent to a remote server, with specific instructions calling *against* the use of their contents. never is a 'repository' updated even once. are these files really diffable? are the repositories really git-natured? could this be a ftp client instead?
Browse files- README.md +44 -0
- adapter_config.json +21 -0
- adapter_model.bin +3 -0
- runs/Aug21_01-07-49_DESKTOP-ALJIUJP/events.out.tfevents.1692605270.DESKTOP-ALJIUJP.1468313.0 +3 -0
- runs/Aug21_01-23-11_DESKTOP-ALJIUJP/events.out.tfevents.1692606191.DESKTOP-ALJIUJP.1468709.0 +3 -0
- training_log.json +14 -0
- training_parameters.json +30 -0
- training_prompt.json +3 -0
README.md
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: peft
|
3 |
+
---
|
4 |
+
## Training procedure
|
5 |
+
|
6 |
+
|
7 |
+
The following `bitsandbytes` quantization config was used during training:
|
8 |
+
- load_in_8bit: True
|
9 |
+
- load_in_4bit: False
|
10 |
+
- llm_int8_threshold: 6.0
|
11 |
+
- llm_int8_skip_modules: None
|
12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
13 |
+
- llm_int8_has_fp16_weight: False
|
14 |
+
- bnb_4bit_quant_type: fp4
|
15 |
+
- bnb_4bit_use_double_quant: False
|
16 |
+
- bnb_4bit_compute_dtype: float32
|
17 |
+
|
18 |
+
The following `bitsandbytes` quantization config was used during training:
|
19 |
+
- load_in_8bit: True
|
20 |
+
- load_in_4bit: False
|
21 |
+
- llm_int8_threshold: 6.0
|
22 |
+
- llm_int8_skip_modules: None
|
23 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
24 |
+
- llm_int8_has_fp16_weight: False
|
25 |
+
- bnb_4bit_quant_type: fp4
|
26 |
+
- bnb_4bit_use_double_quant: False
|
27 |
+
- bnb_4bit_compute_dtype: float32
|
28 |
+
|
29 |
+
The following `bitsandbytes` quantization config was used during training:
|
30 |
+
- load_in_8bit: True
|
31 |
+
- load_in_4bit: False
|
32 |
+
- llm_int8_threshold: 6.0
|
33 |
+
- llm_int8_skip_modules: None
|
34 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
35 |
+
- llm_int8_has_fp16_weight: False
|
36 |
+
- bnb_4bit_quant_type: fp4
|
37 |
+
- bnb_4bit_use_double_quant: False
|
38 |
+
- bnb_4bit_compute_dtype: float32
|
39 |
+
### Framework versions
|
40 |
+
|
41 |
+
- PEFT 0.5.0.dev0
|
42 |
+
- PEFT 0.5.0.dev0
|
43 |
+
|
44 |
+
- PEFT 0.5.0.dev0
|
adapter_config.json
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"auto_mapping": null,
|
3 |
+
"base_model_name_or_path": "models/TheBloke_Llama-2-13B-fp16",
|
4 |
+
"bias": "none",
|
5 |
+
"fan_in_fan_out": false,
|
6 |
+
"inference_mode": true,
|
7 |
+
"init_lora_weights": true,
|
8 |
+
"layers_pattern": null,
|
9 |
+
"layers_to_transform": null,
|
10 |
+
"lora_alpha": 512,
|
11 |
+
"lora_dropout": 0.05,
|
12 |
+
"modules_to_save": null,
|
13 |
+
"peft_type": "LORA",
|
14 |
+
"r": 256,
|
15 |
+
"revision": null,
|
16 |
+
"target_modules": [
|
17 |
+
"q_proj",
|
18 |
+
"v_proj"
|
19 |
+
],
|
20 |
+
"task_type": "CAUSAL_LM"
|
21 |
+
}
|
adapter_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e1ec47ca65429ce6872d0e126a176d568d5d3e69920ade0a06b314fc7d877c9f
|
3 |
+
size 838918733
|
runs/Aug21_01-07-49_DESKTOP-ALJIUJP/events.out.tfevents.1692605270.DESKTOP-ALJIUJP.1468313.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8380dbcdbe8ce53d7b964d82a24917bb3d00fd3a2d2836f2fb2e55c37e2b0476
|
3 |
+
size 4483
|
runs/Aug21_01-23-11_DESKTOP-ALJIUJP/events.out.tfevents.1692606191.DESKTOP-ALJIUJP.1468709.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0c4ff563265baf0237cd2a86000cf0f132669e6ee1c99d13ff12f2f917b5c32
|
3 |
+
size 4831
|
training_log.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"base_model_name": "TheBloke_Llama-2-13B-fp16",
|
3 |
+
"base_model_class": "LlamaForCausalLM",
|
4 |
+
"base_loaded_in_4bit": false,
|
5 |
+
"base_loaded_in_8bit": true,
|
6 |
+
"projections": "q, v",
|
7 |
+
"train_runtime": 4689.4141,
|
8 |
+
"train_samples_per_second": 0.053,
|
9 |
+
"train_steps_per_second": 0.001,
|
10 |
+
"total_flos": 2.311238235193344e+16,
|
11 |
+
"train_loss": 2.563081423441569,
|
12 |
+
"epoch": 0.77,
|
13 |
+
"current_steps": 95
|
14 |
+
}
|
training_parameters.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"lora_name": "greener-13-2-base-8bitD",
|
3 |
+
"always_override": false,
|
4 |
+
"save_steps": 0.0,
|
5 |
+
"micro_batch_size": 2,
|
6 |
+
"batch_size": 64,
|
7 |
+
"epochs": 1.0,
|
8 |
+
"learning_rate": "3e-4",
|
9 |
+
"lr_scheduler_type": "linear",
|
10 |
+
"lora_rank": 256,
|
11 |
+
"lora_alpha": 512,
|
12 |
+
"lora_dropout": 0.05,
|
13 |
+
"cutoff_len": 1536,
|
14 |
+
"dataset": "None",
|
15 |
+
"eval_dataset": "None",
|
16 |
+
"format": "None",
|
17 |
+
"eval_steps": 100.0,
|
18 |
+
"raw_text_file": "greener-pg24246-trim1",
|
19 |
+
"overlap_len": 512,
|
20 |
+
"newline_favor_len": 128,
|
21 |
+
"higher_rank_limit": false,
|
22 |
+
"warmup_steps": 100.0,
|
23 |
+
"optimizer": "adamw_torch",
|
24 |
+
"hard_cut_string": "\\n\\n\\n",
|
25 |
+
"train_only_after": "",
|
26 |
+
"stop_at_loss": 0,
|
27 |
+
"add_eos_token": false,
|
28 |
+
"min_chars": 0.0,
|
29 |
+
"report_to": "tensorboard"
|
30 |
+
}
|
training_prompt.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"template_type": "raw_text"
|
3 |
+
}
|